Wild pollinator communities of native woodlands and commercial almond plantations in a semi-arid Australian landscape:

Implications for conservation of insects and ecosystem services

Manu E Saunders B Arts; B Env Sc (Ecology) Hons, University of Queensland

July 2014

This thesis is submitted in fulfilment of the degree of Doctor of Philosophy Charles Sturt University

Faculty of Science School of Environmental Sciences

Supervisors: Professor Gary W Luck Professor Geoff M Gurr

Dr Margaret M Mayfield (UQ)

ii

iii

iv

v

Table of Contents Page

Preface viii

Acknowledgements ix

Abstract xi

List of Figures xiii

List of Tables xiv

PART I: GENERAL BACKGROUND 1

Chapter 1: Introduction and Scope 3

1.1 Agroecosystem function 3

1.2 Insects as purveyors of ecosystem functions 4

1.3 Aims and Scope 6

1.4 Justification of terms used in this work 8

Chapter 2: Study system and sampling protocol 11

2.1 Study area 11

2.2 Mallee woodlands and shrublands 12

2.3 Almond cultivation in the Mallee 15

2.4 Sampling protocol 16

PART II: ANALYSIS 19

Chapter 3: Almond orchards with living ground cover host more wild

pollinators

23

3.1 Introduction 24

3.2 Methods 27

3.3 Results 37

3.4 Discussion 44

Chapter 4: Wild pollinators in flowering almond orchards and

proximity to unmanaged vegetation

49

4.1 Introduction 50

4.2 Methods 52

4.3 Results 59

4.4 Discussion 65

Chapter 5: The influence of a semi-arid woodland—almond

plantation edge on wild pollinator communities

71

5.1 Introduction 72

vi

5.2 Methods 75

5.3 Results 81

5.4 Discussion 88

Chapter 6: Food and habitat resources available to wild pollinators in

a late winter flowering crop

95

6.1 Introduction 96

6.2 Methods 99

6.3 Results 104

6.4 Discussion 114

PART III: SYNTHESIS 121

Chapter 7: Synthesis and Conclusions 123

7.1 Synthesis and general discussion 123

7.2 Conclusions 126

PART IV: REFERENCE MATERIAL 129

Literature Cited 130

Appendix 1: Pan trap catches of pollinator insects vary with habitat

(published manuscript)

163

Appendix 2: Taxonomic lists of flower visitor observations and trap

samples

183

Appendix 3: Full set of models for analyses of ground cover vegetation—

pollinator relationships (Chapter 3)

197

Appendix 4: Location attributes and pollinator occurrence along ecotone

gradients (Chapter 5)

202

Appendix 5: Physical attributes and full candidate sets of models for

analysis of habitat structure relationships (Chapter 6)

211

vii

Certificate of Authorship

I hereby declare that this submission is my own work and to the best of my

knowledge and belief, understand that it contains no material previously published

or written by another person, nor material which to a substantial extent has been

accepted for the award of any other degree or diploma at Charles Sturt University

or any other educational institution, except where due acknowledgement is made

in the thesis. Any contribution made to the research by colleagues with whom I

have worked at Charles Sturt University or elsewhere during my candidature is

fully acknowledged. I agree that this thesis be accessible for the purpose of study

and research in accordance with normal conditions established by the Executive

Director, Library Services, Charles Sturt University or nominee, for the care, loan

and reproduction of thesis, subject to confidentiality provisions as approved by the

University.

Name:

Signature:

Date:

viii

Preface

This thesis includes five individual manuscripts that have been published in academic journals. Consequently, there will be some repetition across chapters, especially within discussions of background material and methodological details. Chapter 2 provides a brief summary of the study area and an overview of the general sampling approach but specific methodological and analytical procedures for each individual study are detailed in each chapter. A single reference list is provided at the end of the thesis to avoid unnecessary duplication. I am personally responsible for conceptualising the project, all data collection and analysis, and the majority of written text. Gary Luck, Margie Mayfield and Geoff Gurr contributed to the planning and sampling design stages and to the papers they have co-authored with me.

Peer-reviewed publications from this work:

Chapter 3: Almond orchards with living ground cover host more wild insect pollinators. (Saunders ME, Luck GW, Mayfield MM (2013) Journal of Insect Conservation 17:1011-1025).

Chapter 5: Spatial and temporal variation in pollinator community structure relative to a woodland—almond plantation boundary. (Saunders ME, Luck GW (2014) Agricultural and Forest Entomology, doi:10.1111/afe.12067).

Chapter 6: Keystone resources available to wild pollinators in a winter-flowering tree crop plantation. (Saunders ME, Luck GW, Gurr GG (in press) Agricultural and Forest Entomology).

Appendix 1: Pan trap catches of pollinator insects vary with habitat. (Saunders ME, Luck GW (2013) Australian Journal of Entomology52:106-113).

Conference presentations from this work:

Saunders ME, Luck GW, Mayfield MM (2012) Living ground cover influences native pollinator abundance in commercial almond orchards. 1st

Apimondia ApiEcoFlora Symposium, San Marino, Republic of San Marino, 4-6 October 2012, poster presentation. (Chapter 3)

Saunders ME, Luck GW (2013) Pan trap catches of pollinator insects vary with habitat context. VII Southern Connection Congress, Dunedin, New Zealand, 21-25 January 2013, poster presentation. (Appendix 1)

ix

Acknowledgements

Submission of this thesis is a beginning for me, not an end. I am especially

grateful to my principal supervisor, Gary Luck, who provided the precise relative

proportions of humour, guidance, encouragement and patience that I needed. He

gave me freedom to learn to be the ecologist I am, and direction when tangents

knocked at my door. I could not have asked for a better supervisor. My co-

supervisor, Margie Mayfield, has been an equally inspirational force, before I

even knew I was going to undertake a PhD. I cannot thank her enough for giving

me the opportunity as an undergraduate student to develop my interest in

pollination ecology and for being part of this project. I also sincerely thank Geoff

Gurr, for project advice and intellectual contributions; the Institute for Land Water

and Society and the School of Environmental Sciences, for financial,

administrative and logistic support; Simon McDonald, Deanna Duffy and Gail

Fuller, for technical support, patience with all my ignorant questions about GIS

and GPS and swiftness in delivering top-quality conference posters; the

indispensable lab staff, particularly Lauren Sheather, who helped with equipment,

field work preparation, and assistance during the long days I spent staring down a

microscope; Ken Walker for voluntary identification of my (very poorly-

mounted) native bee specimens; Select Harvests and Manna Farms, for kindly

providing access to their almond trees; and the Department of Sustainability and

Environment, Victoria, for research access to mallee national parks. Many other

people have contributed, either directly or indirectly, to me reaching this point (in

no particular order): Alison Matthews, Lisa Smallbone, Paul Humphries, Nicole

McCasker, Ian Lunt, Gill Earl, Janet Cohn, Katrina Lumb, Tobias Smith, Peter

Spooner, Wayne Robinson, Antonio Castro, Simon Watson, Saul Cunningham,

Alan Yen, Michael Batley, Peter Lillywhite, Gimme Walter and Barbara

Gemmill-Herren. James, you and Sarge have kept me afloat over the last 3.5

years, and there would have been more tears than smiles without you and the

music. Most importantly, I thank my mother for teaching me to believe in what

truly matters.

x

xi

Abstract

Understanding how monocultures of introduced crop species interact with native

ecosystems and wildlife in the surrounding landscape is vital for the ecological

management of agroecosystems, and can have far-reaching benefits for

biodiversity conservation and crop yields. Insects make essential contributions to

ecosystem functions in all types of ecosystems and many functions are

synergistically linked, like pollination and biological pest control. Thus, the

conservation of invertebrate communities in agricultural landscapes is paramount.

Although many studies have investigated pollinator communities using field crops

(e.g. canola), more information is needed on how plantations of deciduous tree

crops influence local insect communities, as these agroecosystems can create

permanent spatial and temporal homogeneity across agricultural landscapes.

I sampled potential wild pollinators in commercial almond plantations and

native mallee vegetation in northwest Victoria, Australia, over two almond

flowering seasons in late winter of 2010 and 2011. The aim of the study was to

provide information on the abundance, richness and composition of insect

pollinator communities at this time, which is the critical life stage for both almond

trees (reproduction) and insects (emergence and reproduction). In particular, I

focused on comparing patterns of insect community distribution relative to

environmental attributes of mallee vegetation and almond plantations, with the

goal of identifying how plantations might influence the conservation of wild

insect communities, and thus enhance ecosystem service provision in commercial

plantations.

Overall, I found that potential wild pollinators were using almond

plantations in the Victorian mallee during the almond flowering season. However,

their movement into plantation interiors was limited by homogeneous vegetation

structure and a lack of essential resources, such as temporal continuity of food

resources and the availability of nesting sites. Wild pollinators, particularly native

bees, were more abundant in small almond orchards with weedy ground cover

across orchard floors, compared to both monoculture plantations and native

mallee vegetation. In small orchards surrounded by a heterogeneous agricultural

mosaic landscape, abundance and richness of pollinator groups were negatively

associated with the proximity of native vegetation. However, in broadacre

xii

monoculture plantations, most pollinator taxa were limited to edges near native

mallee vegetation, or within remnant patches of mallee in the plantation interior.

In particular, native bees were only collected in remnant mallee patches inside the

plantations, rather than in blocks of almond trees. Before and after almond

flowering, when resources inside plantations were limited, pollinator groups were

mostly associated with sites that provided alternative foraging and nesting

resources, such as leaf litter, ground cover vegetation or the presence of

herbivorous insect prey.

The work presented in this thesis has two-fold significance. My results

provide valuable information on potential wild pollinators using Australian

almond plantations, which has the capacity to motivate the adoption of ecological

management goals among growers and enhance awareness of the essential

contribution of wild insects to ecosystem functions. I also highlight important

knowledge gaps in our understanding of the diverse invertebrate fauna of the

Victorian mallee regions, an area of high conservation value, which can inform

future research and conservation goals.

xiii

List of Figures

Figure Caption Page2.1 Map of the study region and sampling localities 112.2 Wild insect pollinators of the Victorian mallee 132.3 Typical mallee plant communities of the study region 143.1 Representative photographs of the three sampled vegetation

types: plant-diverse orchards, monoculture orchards, native mallee vegetation

30

3.2 Species accumulation curves for insect pollinator groups in each sampled vegetation type

39

3.3 Diversity profiles for insect pollinator groups 413.4 Dendrograms showing pollinator community similarity between

each site group43

4.1 Representative images of landscape composition in Complex and Simple landscapes

53

4.2 Ordination plot showing overall pollinator community composition between landscape types

58

4.3 Ordination plots showing pollinator community similarity between sites of difference resource profiles in each landscape type

60

4.4 Proportional abundance and richness of each pollinator group in each landscape type

61

4.5 Proportional abundance of pollinator groups collected in almond blocks and interior mallee patches in Simple landscape

64

5.1 Map showing the almond/mallee edges sampled for ecotone study

76

5.2 Representative photographs showing phenological changes occurring across plantations during the flowering season

77

5.3 Distance-decay graph for decline in pollinator community similarity with distance

82

5.4 Average pollinator group abundance and richness for each month at mallee and almond sites

84

5.5 Monthly turnover of wasp and fly abundance and richness along the ecotone gradients

85

6.1 Pollinator community similarity between mallee and almond sites 106A1.1 Map of the study area 166A1.2 Proportional abundance of total insects collected in each pan trap

colour172

A1.3 Proportional abundance of total native Hymenoptera collected in each pan trap colour

172

A2.1 Representative photographs of insect pollinators observed visiting flowers in mallee vegetation and almond plantations

184

xiv

List of Tables

Table Caption Page3.1 Almond orchards and mallee sites used for sampling 283.2 Average abundance and richness of insect pollinators in each

vegetation type38

3.3 Parameter estimates for relationships between pollinators and ground cover attributes

42

4.1 Abundance and richness of pollinator groups and proportion of land use types around sites in each landscape type

57

4.2 Relationships between pollinator groups and proportion of unmanaged vegetation types around sites

64

5.1 SHE analysis: comparison of physical and taxonomic boundaries along ecotone gradient

87

6.1 Average pollinator abundance and richness and resource attributes at mallee and almond sites

105

6.2 Relationships between pollinator groups and overall site heterogeneity

108

6.3 Relationships between pollinator groups and resource attributes 1096.4 Summed Akaike weights for each predictor variable (resource

attributes)113

A1.1 Total abundance of pollinator groups collected in each colour pan trap

170

A1.2 Results of Marascuilo procedure comparing proportion of pollinators collected in each colour pan trap

171

A2.1 Observations of flower visitors visiting almond and native flowers 183A2.2 Taxonomic list of samples 2010 185A2.3 Taxonomic list of samples 2011 192A3.1 Full candidate model set for relationships between pollinator

insects and ground cover vegetation197

A4.1 Physical attributes of sampled patches in ecotone study 202A4.2 Occurrence of Diptera and Hymenoptera family groups along

ecotone gradients203

A5.1 Physical attributes of almond plantations in resource study 211A5.2 Full list of candidate models for pollinator relationships with

resource attributes212

1

PART I: GENERAL BACKGROUND

2

3

Chapter 1: Introduction and scope

1.1 Agroecosystem function

Agroecosystems provide a foundation for human civilisation and incorporating

knowledge of ecosystem functions and services into their management is

essential. Any system that is structured around living organisms will respond to

natural cycles and processes in the absence of human inputs. Thus, ensuring the

long-term productivity of agroecosystems depends on understanding how they

function as a system and how they interact with the natural ecosystems and

landscapes in which they are embedded (Pimentel et al. 1989). A useful way to

investigate the ecology of agroecosystems is to examine communities of

ecosystem service providers, such as pollinating insects. Land use change often

impacts processes of community assembly more than individual species per se

(Mayfield et al. 2010; Cadotte et al. 2011) and the delivery of most ecosystem

services is synergistically linked across multiple species and interactions (Luck et

al. 2009; Macfadyen et al. 2012; Lundin et al. 2013).

Every process or exchange of energy that occurs between an ecosystem’s

components is a ‘function’ of that ecosystem (Şekercioğlu 2010) and the

ecological outcomes of these functions generally provide beneficial ‘services’ to

humans (Vogt 1948; Ehrlich & Mooney 1983; Daily 1997). A greater diversity of

species, community patterns and natural processes allows for greater availability

of ecosystem services (Altieri 1999; Loreau et al. 2001; Symstad et al. 2003;

Mayfield et al. 2010; Maestre et al. 2012). However, biotic diversity is often

lacking in agroecosystems. Most modern agroecosystems aim to ‘standardise’

production by disconnecting from ecological cycles through the segregation of

animals and plants, isolation from local ecological knowledge, and cultivation of

plant monocultures (Weis 2010; Woodhouse 2010). In some parts of the world,

landscape-scale heterogeneity and ecological complexity are being rapidly

replaced by extensive vegetation homogeneity in the form of broadacre crop

monocultures (Hartemink 2005; Plourde et al. 2013). Yet, natural systems are

never wholly autonomous or static, either in configuration or function (Odum

1969; Ulanowicz 1990; Bejan & Lorente 2010), so intensively-managed

homogeneous agroecosystems contravene the principles of natural systems and

4

can disrupt local availability of ecosystem services. Thus, there is a need for more

studies that identify ecological and structural differences between crop

monocultures and local native vegetation, and relate these differences to the

distribution of ecological communities (e.g. Ramirez & Simonetti 2011;

Dąbrowska-Prot & Wasiłowska 2012; Meng et al. 2012; Pryke & Samways

2012). Results from such studies would provide valuable information for

biodiversity conservation and the ecological management of agroecosystems,

which can assist managers to balance long-term productivity with a reduction in

input costs.

1.2 Insects as purveyors of ecosystem services

Many ecosystem services that are beneficial to humans are a result of processes or

interactions to which invertebrates are key contributors. These include pollination,

biological control, soil formation and nutrient cycling. Mutualisms, such as plant-

pollinator interactions, are particularly critical to ecosystems and there are far-

reaching consequences if these interactions are lost (Kiers et al. 2010). Pollinators

indirectly affect many other parts of ecosystems and their loss could result in

significant changes to both the physical appearance of the ecosystem and its

biogeochemical structure through changes in plant community composition

(Ashman et al. 2004; Knight et al. 2005; Biesmeijer et al. 2006; Pauw 2007; Pauw

& Hawkins 2011) and, subsequently, soil structure and the hydrologic cycle

(D’Odorico et al. 2010). All ecosystem services are interdependent (Daily 1997;

Kremen et al. 2007; Bennett et al. 2009) and recent evidence has shown that other

services important to agroecosystems, such as biological pest control, can be

synergistically linked to the provision of pollination services (Bos et al. 2007;

Lundin et al. 2013). Thus, pollinator communities are an ideal focal group to

study how agroecosystems influence local ecological communities.

Plantations of perennial tree crops, such as orchard fruits, may have

particular influence over native pollinator communities in agricultural landscapes.

Many deciduous fruit trees are pollinator-dependent (Kevan 1999), but if they are

managed as intensive monoculture plantations they create large areas of spatial

and temporal homogeneity (in vegetation structure and resource availability) that

are inhospitable to pollinators before and after the brief flowering period

5

(Sheffield et al. 2008; Marini et al. 2012). Compared to field crops, which are

most commonly grown in broadacre monocultures, tree crop plantations can have

a very different ecological influence over the surrounding ecosystems. Field

crops, such as wheat, canola or maize, are grown in short rotation cycles and

alternated with other field crop varieties, which may contribute temporal

heterogeneity to the landscape that could counteract the spatial homogeneity

(Vasseur et al. 2013). In contrast, tree crop plantations are embedded in the

landscape, often for decades, creating an ecologically stable, semi-permanent

ecosystem (Tedders 1983; Altieri 1999). This could potentially influence the

structure of local invertebrate communities in the long-term (Harding et al. 1998;

Tscharntke et al. 2002). The conservation value of tree crop agriculture has long

been recognised (Smith 1916; Altieri et al. 1983; Ryszkowski & Kędziora 1987)

but the environmental influence of tree plantations in agricultural landscapes has

historically received less attention in the scientific literature than annual crops

(Hartemink 2005). In particular, most of the published studies investigating the

influence of crop monocultures on pollinator taxa or communities have focused

on field crops, such as canola, legumes or vegetables (e.g. Westphal et al. 2003;

Gemmill-Herren & Ochieng 2008; Arthur et al. 2010; Rader et al. 2012). These

studies provide valuable insight into how agricultural land uses affect pollinators,

but tree monocultures are very different in structure and phenology to

monocultures of herbaceous or shrubby species and are likely to influence

pollinator communities, and the surrounding environment, in different ways.

Some important pollinator taxa, particularly bees and hoverflies, often

prefer foraging in relatively open areas dominated by early successional

vegetation (Winfree et al. 2011; Robinson & Lundholm 2012) compared to dense

forest systems (Gittings et al. 2006; Hoehn et al. 2010; Proctor et al. 2012).

Intensively-managed tree crop plantations can be structurally similar to dense

forests, and this could affect the availability of pollinator communities for tree

crop varieties that rely on pollinators to set fruit. Although some studies have

provided valuable information on the effects of tree plantations on local insect

communities, these have largely been relevant to plantation forests managed for

timber production (e.g. Humphrey et al. 1999; Gittings et al. 2006; Campbell et al.

2011; Pryke & Samways 2012). Pollinator-dependent orchard fruit plantations

(e.g. apple, almond, stone fruit) are managed very differently to timber forests,

6

particularly with regard to orchard floor vegetation and disturbance cycles. Given

the recent increase in global demand for orchard fruit production (McKenna &

Murray 2002; Jackson et al. 2011), understanding how these systems influence

local flora and fauna is paramount, particularly in Australia where the majority of

orchard crop varieties are exotic species (Cunningham et al. 2002). Previous work

on pollinators in Australian tree plantations has been conducted in plantations of

native macadamia or Eucalyptus species (e.g. Heard & Exley 1994; Barbour et al.

2008) and evergreen tropical tree crops like cashew or longan (e.g. Heard et al.

1990; Blanche et al. 2006). This thesis aims to address some important knowledge

gaps by presenting the first study of wild pollinator insects using monoculture

plantations of introduced almond trees, a pollinator-dependent deciduous tree

crop, in a semi-arid region of southern Australia.

1.3 Aims and scope

The research presented here compares abundance, richness and distribution of

potential wild pollinator insects in almond plantations of north-west Victoria with

pollinator communities in native mallee vegetation. Furthermore, it intends to

provide insight into how intensive monoculture plantations of orchard fruit trees

interact with the landscape, ecosystems and biotic communities around them. To

do this, I asked the following questions:

1. Are wild pollinator communities influenced by the presence of living

ground cover under almond trees? (Chapter 3)

2. Is there a relationship between wild pollinator abundance and diversity in

almond orchards and the proportion of remnant native vegetation in the

surrounding matrix? (Chapter 4)

3. How does wild pollinator abundance and diversity change across habitat

boundaries between almond monocultures and adjacent native mallee

vegetation? (Chapter 5)

4. Do winter-flowering monoculture plantations provide food and nesting

resources for overwintering and emerging wild pollinator taxa? (Chapter

6)

My underlying hypothesis is that wild pollinator communities will be more

species rich and abundant, and therefore more available to pollinate almond

7

flowers, at sites with a greater diversity of food and habitat resources. My

sampling approach focused on the late winter flowering season, as this is a critical

life stage for almond trees (flowering and nut formation) and pollinator insects

(emergence and reproduction). Hence, supporting the conservation of wild

pollinator communities in the long-term will depend on matching habitat and

resource requirements of pollinators with the structural and phenological

attributes of the plantation at this time. The structure of this thesis is as follows:

Chapter 2 presents a brief summary of the ecological significance of the

study area and discusses the general sampling protocol;

Chapter 3 investigates relationships between pollinator groups and weedy

ground cover vegetation on almond orchard floors. This provides

important information for almond growers in Australia, many of whom

actively remove vegetation on plantation floors to increase management

efficiency, despite the important benefits ground cover provides, such as

erosion control, improved soil structure and habitat for beneficial insects

(Eilers & Klein 2009; Ramos et al. 2010);

Chapter 4 investigates whether the composition of pollinator communities

in orchards during flowering is associated with the presence of natural or

semi-natural vegetation near sites. This knowledge has implications for

almond management and local biodiversity conservation, because a matrix

composed of a diverse array of food and habitat resources is essential to

support pollinators in agricultural landscapes dominated by perennial

crops that flower for a brief period once a year (Williams & Kremen 2007;

Sheffield et al. 2008);

Chapter 5 examines how the physical habitat boundary between mallee

woodlands and almond plantations influences the distribution of pollinator

communities before, during and after the brief almond flowering event.

This contributes to ecological knowledge of how insects might respond to

abrupt changes in spatial and temporal distribution of resources in

agroecosystems;

Chapter 6 examines the potential for late-winter flowering monoculture

plantations to support overwintering or emerging pollinator insects. This

has rarely been considered in studies of agroecosystem impacts on

pollinator species, and provides valuable insights into how pollinator

8

communities use plantations of almond trees during their emergence and

reproductive stages, which coincides with the almond reproductive stage;

Chapter 7 provides a brief synthesis and discussion of the overall results.

In particular, I link results from individual studies to highlight the overall

significance and contribution of this thesis, and provide recommendations

for future research and management goals.

Finally, the Appendices provide additional information relevant to the

overall thesis. In particular, Appendix 1 presents the manuscript of an

original publication based on data collected during my PhD research,

which contributes to the methodological literature by discussing important

considerations for designing ecologically meaningful pan trap studies; and

Appendix 2 lists the total abundance of bee species and wasp and fly

family groups collected across both sampling seasons.

1.4 Justification of terms used in this work

‘Pollinator’ refers to individuals in the taxonomic orders Diptera (flies)

and Hymenoptera (bees, wasps). Other insect groups are also common

pollinators, like Lepidoptera (moths and butterflies), Thysanoptera

(thrips), and Coleoptera (beetles), however my sampling method (see

Chapter 2) was designed to target Hymenoptera and Diptera individuals,

as these are the most commonly-recorded visitors to open blossoms of

almond trees and related species (Rosaceae: Prunus spp.) in other parts of

the world (e.g. Abrol 1988; Jordano 1993; Bosch & Blas 1994). Although

I did not quantify pollination efficiency in this study, I use the term

pollinator to maintain consistency with the crop pollination literature

because one of the main aims of my research was to document wild insect

species using almond orchards that had the potential to provide pollination

services to growers. Flies and wasps are important, but understudied,

pollinators in a variety of natural and agricultural systems (Kevan & Baker

1983; Larson et al. 2001; Ssymank et al. 2008), particularly in native

Australian ecosystems (Armstrong 1979; Pickering & Stock 2003). Many

pollinating Diptera and non-bee Hymenoptera are also economically-

important predators and parasitoids of herbivorous pests. Because

9

ecosystem service provision often involves multiple species and their

interactions (Luck et al. 2009; Lundin et al. 2013), studies of ecosystem

services benefit from concentrating on groups of taxa that provide the

same services (Bennett et al. 2009; Macfadyen et al. 2012), rather than on

a single, well-known taxon (e.g. bees).

‘Broadacre monoculture’ refers to large-scale, intensive cultivations of a

single crop species covering more than 100 hectares of land. This type of

agroecosystem is more common in developed countries with substantial

areas of rangeland, like Australia, Canada, and the United States of

America and is mostly used to cultivate fast-growing, high-yielding crop

varieties that require high chemical usage and seed purchase for each

planting (Mason 2003; Plourde et al. 2013). More recently, this practice

has become common in commercially-grown perennial tree crop or

plantation crop systems, such as oil palm, orchard crops and agroforestry

schemes (Hartemink 2005).

‘Orchard’ and ‘plantation’ are used interchangeably in this work to refer to

a commercial almond cultivation system. The term ‘orchard’ has mainly

been used when comparing small plant-diverse orchard systems with

broadacre monoculture plantations (Chapters 3-4), as the most descriptive

term relevant to both types of cultivation system was chosen to maintain

consistency throughout individual papers. However, the goal of this thesis

was to investigate and discuss the ecological role of monoculture

plantations, using intensive almond plantations as a study system; hence,

the term ‘plantation’ is used elsewhere.

10

11

Chapter 2: Study system and sampling protocol

2.1 Study area

The study area is in north-west Victoria, Australia, within the Murray Mallee

bioregion (Department of Sustainability and Environment 2013). Sampling sites

for all studies were within native mallee vegetation and commercial almond

plantations located in the districts of Wemen, Liparoo, Colignan and Nangiloc

(Figure 2.1). The region is semi-arid with a Mediterranean climate. Annual

rainfall varies widely with an average of around 300 mm per year. The region was

wetter than average during the sampling years, with approximately 528 mm

falling in 2010 and approximately 566 mm falling in 2011 (Bureau of

Meteorology 2013). Average temperatures for the months of sampling (July—

September) ranged from minima of about 5o C to maxima of about 20o C.

Figure 2.1 Map of the study region showing localities of sampled almond plantations.

12

2.2 Mallee woodlands and shrublands

The term ‘mallee’ was originally used to describe the open scrub growing on

sandy plains of north-west Victoria and eastern South Australia, which was

dominated by Eucalyptus dumosa, or white mallee (Westgarth 1848). Today,

‘mallee’ refers to semi-arid open heath and woodlands native to Mediterranean

climate zones in southern Australia, specifically the southern coast of Western

Australia, parts of South Australia, north-west Victoria and western New South

Wales. Mallee ecosystems occur in regions receiving between 250 and 500 mm of

average annual rainfall and are characterised by sandy dune systems and

predominantly alkaline soils, often with a shallow surface layer overlying

limestone bedrock (Noble & Bradstock 1989). Open woodlands are dominated by

Eucalyptus species growing in the coppiced ‘mallee’ habit of shorter trees with

multiple stems of similar age developing from a large subterranean lignotuber, as

well as Acacia, Callitris and Casuarina species (Wood 1929). The understorey is

open and highly variable, with patchy shrub and ground cover vegetation typically

composed of saltbush and other halophytic plants, Triodia hummock grass,

cryptogams, grasses and flowering forbs, and including a high number of rare

species (Whittaker et al. 1979; Blakers & Macmillan 1988; Cheal et al. 2013). In

addition, the mallee ground layer is characterised by mosaics of bare ground and

dense mats of leaf litter and decaying wood (Holland 1969; Whittaker et al. 1979).

Knowledge of pollinator insect species (Figure 2.2) and plant-pollinator

relationships in the study region is limited (A.L. Yen, pers. comm., 6 July 2011).

A search of the peer-reviewed literature (ISI Web of Knowledge, 1 May 2014)

produced 11 published ecological studies on invertebrate fauna in the Victorian

mallee – two of these were from the current thesis (Saunders & Luck 2013;

Saunders et al. 2013; Chapter 2 and Appendix 1), five studied ant communities

(e.g. Andersen 1983; Andersen & Yen 1992), and three studied introduced

European honeybees (e.g. Oldroyd et al. 1994; Horskins & Turner 1999). Yen et

al. (2006) provide an observational account of native bees, and suggest that native

bee abundance in the Victorian mallee was high and estimating that the area could

support about 10 % of Australian bee species, which concurs with Michener’s

(1979) conclusion that bees are most abundant in warm, temperate, xeric areas,

such as the semi-arid mallee regions. A number of empirical and observational

13

studies have included accounts of various species of native Diptera and non-ant

Hymenoptera visiting mallee flowers (e.g. Ford & Forde 1976; Alcock 1981;

Hawkeswood 1982; Horskins & Turner 1999; Yen et al. 2006). Stephens et al.

(2006) also found that a high proportion of wasp species collected in a South

Australian mallee ecosystem were only collected on a single plant species. Like

most semi-arid ecosystems, the mallee is highly influenced by fire and rainfall

events, although little information is available on how these events affect

pollinator communities in the mallee. Mallee ecosystems support a high

proportion of leaf litter and dead wood, which are important habitats for many

invertebrate species (Austin et al. 2005; New et al. 2010). Hence, intense wildfires

could severely diminish local insect communities (New et al. 2010), or the

characteristic mosaic distribution of these attributes could increase the availability

of unburned refuge sites during patchy fire events (Majer et al. 1997; Andrew et

al. 2000). Andersen & Yen (1985) found that wildfire had immediate effects on

ant communities in the Victorian mallee, including a shift toward greater

compositional evenness; however, Friend & Williams (1996) studied the long-

term effects of fire in a Western Australian mallee ecosystem and found that

variation in native bee, wasp and fly communities appeared to be more influenced

by general seasonal effects than fire-related trends.

Figure 2.2 Wild insect pollinators of the Victorian mallee: (left) A hoverfly (Diptera: Syrphidae) resting on weedy Brassicaceae spp. in a Wemen mallee remnant; (middle) A bee fly (Diptera: Bombyliidae) resting on leaf litter, Murray-Kulkyne National Park; (right) Tiphiid wasps (Hymenoptera: Tiphiidae) in copula on a Leptospermum coriaceum flower, Hattah-Kulkyne National Park (see also Appendix 2).

Australian mallee regions provide ideal habitat for many pollinator insect species

(Figure 2.3). Most pollinator taxa, particularly solitary and eusocial bees, rely on

14

multiple proximate habitats that provide different, but complementary food and

habitat resources needed throughout their life cycle – this has been called

‘landscape complementation’ by Dunning et al. (1992), the problem of ‘partial

habitats’ by Westrich (1996) and ‘functional connectivity’ by Williams & Kremen

(2007). Mild winters, low mallee tree canopies (Andersen & Yen 1992), diverse

floral resources, particularly in the herb and shrub layers (Yen et al. 2006), and

open vegetation structure create an optimal foraging environment for pollinator

insects (Kremen et al. 2007; Winfree et al. 2011). In addition, the patchy

distribution of bare, sandy soil and leaf litter banks are preferred substrates for

many species of overwintering and ground-nesting bees, flies and wasps (Knerer

& Schwarz 1976; Leather et al. 1993; Potts et al. 2005; Houston & Maynard

2012).

Figure 2.3 Typical mallee plant communities of the study region: (top) mallee woodland interior and (bottom) mallee heath and shrubland, Annuello State Flora and Fauna Reserve.

15

2.3 Almond Cultivation in the Mallee

Large-scale clearing and cultivation began in the Victorian mallee in the early

1900s after failed attempts to establish a pastoral industry, and the region has

subsequently developed into one of Australia’s main irrigated agricultural regions,

the Sunraysia district, relying on intensive water extraction from the Murray River

(Duncan & Dorrough 2009). Agriculture in the mallee region is environmentally

risky (Sadras et al. 2003) because the highly erratic rainfall, poor soil fertility, and

fast evaporation and infiltration rates mean that surface water is rare (Blakers &

Macmillan 1988). Since European settlement, native mallee flora and fauna

communities have been severely affected by clearing and fragmentation

(Cunningham 2000; Driscoll & Weir 2005), rabbit and mouse plagues (Onans &

Parsons 1980; Singleton et al. 2001), and increased soil salinity and groundwater

salt leakage (Cole et al. 1989).

Almond (Rosaceae: Prunus dulcis Mill.) production has been particularly

successful in the Victorian mallee, with the region supporting 64 % of the

country’s almond plantings (Almond Board of Australia 2013). Almond

originated in the foothills and desert areas of central and southwest Asia Minor,

and is adapted to mild, wet winters and hot, dry summers (Kester & Ross 1996;

Ladizinsky 1999), typical of Australia’s mallee regions. The first small-scale

commercial almond plantings were in South Australia in the late 1800s

(Wilkinson 2012) and Australia is now the second-largest almond producer in the

world, with almonds contributing more to the total annual nut production than the

native macadamia (Almond Board of Australia 2013). Almond trees require 100%

cross-pollination to set fruit (Hill et al. 1985; Weinbaum 1985), which means that

managed European honeybee (Apis mellifera) hives have become an essential, and

expensive, part of production for commercial almond growers (vanEngelsdorp et

al. 2008). This reliance on commercial honeybees raises serious concerns for

Australian almond production as impacts such as Varroa mite, climate change,

and agricultural intensification increasingly affect the global honeybee industry

(vanEngelsdorp et al. 2008; Spivak et al. 2011; Henry et al. 2012). Yet, only a

handful of studies have investigated almond pollination in Australia (Hill et al.

1985; Rattigan & Hill 1987; Vezvaei 1994; Vezvaei & Jackson 1995; Jackson

16

1996), and these have looked at the reproductive biology, gene flow or phenology

of the flowers and make little reference to pollinating organisms.

A particular challenge in managing almond pollination services is that

individual flowers are receptive to pollination for an extremely brief period

(approximately 4-5 days) in late winter (Griggs & Iwakiri 1975; Hill et al. 1985),

when most insect activity is limited. Because of this, it is likely that many growers

have assumed European honeybees are the only pollinator option (Klein et al.

2012). However, Mediterranean plant communities in other parts of the world

typically support a high number of insect-pollinated species, and pollinator taxa

and flowering plants are both common (although less abundant) during cooler

months (Petanidou & Vokou 1990; Dafni 1996; Bosch et al. 1997). In fact, despite

many individual plant species having specific flowering periods, the

complementarity of each one’s phenology can provide the temporal resource

connectivity needed to sustain some pollinator taxa, particularly Diptera,

throughout the year (Petanidou & Lamborn 2005).

2.4 Sampling Protocol

Potential wild pollinators were sampled in almond plantations and native mallee

vegetation during the almond flowering seasons between July and September,

2010 and 2011. In 2010, sites were located in broadacre monoculture plantations

at Wemen and Liparoo (n = 30), small, plant-diverse orchards at Colignan and

Nangiloc (n = 27) and native mallee vegetation (n = 45) (Chapters 3 and 4). In

2011, sampling focused on broadacre monoculture plantations (n = 50) and

adjacent native mallee vegetation (n = 18) at Wemen and Liparoo (Chapters 5 and

6). Site selection on the 2010 sampling trip was limited by access to farms in the

region that maintained living ground cover across orchard floors. New South

Wales government extension advice recommends that growers maintain bare soil

orchard floors to minimise frost damage and competition for soil moisture (NSW

Agriculture 2002; Wilkinson 2012). Hence, most commercial almond growers in

the study region use intensive mowing and herbicide application to control

herbaceous ground cover throughout plantations. Monoculture plantations and

plant-diverse orchards were selected to maximise the similarity between sites of

each management type. All plant-diverse orchards at Colignan and Nangiloc were

17

owned and managed by the same family-owned company (Manna Farms). Two of

these orchards were biodynamically-managed (a type of intensive organic

management) and one was managed as a low-intensity conventional farm with no

chemical use (see Chapter 3 for details). The two monoculture plantations were

owned and managed by the same corporation (Select Harvests) and management

practices were identical at both plantations. No insecticides were used in these

plantations, but limited application of herbicide or fungicide was used at specific

times of the year (see Chapter 3). Chemical applications did not occur during

sampling. In both years, site selection was also limited by logistics and weather

conditions, due to widespread flooding and above-average rainfall across much of

the study region during the sampling years. Low-lying areas of plantations were

flooded and access to some parts of the national parks used for native vegetation

sites was also limited, particularly during the 2011 sampling trips. Native

vegetation sites were located in the Annuello State Flora and Fauna Reserve, the

Murray-Kulkyne National Park (Department of Sustainability and Environment,

Victoria: Research Permit Number 10005774) and various ungazetted patches of

remnant mallee near almond blocks.

Insects were collected at each site using sets of pan traps (2010: blue,

yellow and white; 2011: yellow and white). Traps were made from plastic picnic

bowls (Woolworths Home Brand, Bella Vista NSW) painted with UV-bright

fluorescent paint (White Knight Paints, Lansvale NSW) or left white (dimensions:

18 cm diameter, 3 cm depth). Pan traps can underestimate the full composition of

insect assemblages; however this method was used because the priority was to

maximise sampling time per site and minimise collector bias (Westphal et al.

2008). To increase the likelihood that traps provide ecologically meaningful

results, trap colour and placement should be designed to suit the context of the

study system and research questions (Tuell & Isaacs 2009; Vrdoljak & Samways

2012; Saunders & Luck 2013; see Appendix 1). Hence, yellow and white were

chosen because these were the most common flower colours at my sites (in mallee

and almond respectively). In 2010, blue traps were also included as this colour is

commonly thought to attract bees (usually European honeybees), although blue

traps have not been widely tested for attracting native bees in Australia (see

Appendix 1 for discussion of this). In 2011, only yellow and white were used,

based on the results from 2010 sampling that indicated the target insects (non-Apis

18

bees, wasps and flies) were captured more frequently in yellow and white traps

(Appendix 1).

At each site, pan traps were placed together on the ground about 1 m from

a flowering tree to ensure maximum visibility to flying insects. Ground placement

was also chosen because the late winter sampling period meant that some insects

would be emerging from pupae or overwintering sites, and many of my target

species overwinter or nest in the ground or in leaf litter and ground cover

vegetation (see Chapter 6). Each site was sampled once per sampling trip; traps

were set out for 8-10 hours during the warmest part of the day to capture

maximum insect activity and were collected in the sequence they were set out. In

2010, flower visitors on open flowers were also recorded as baseline observations,

but these data have not been used in analyses due to the low individual counts.

Visual surveys were conducted at almond and mallee trap sites by random walks

within 10 metres of trap locations for a 10 minute period. All insects seen visiting

open flowers were counted and identified to the highest possible taxonomic level

(see Appendix 2 for data and photographs).

Upon collection from pan traps, insects were stored in 70 % ethanol and

counted and sorted in the Charles Sturt University Ecology laboratory using an

Olympus SZ51 microscope. The full collection, including voucher specimens, was

stored at Charles Sturt University laboratory. Because of the limited ecological

and taxonomic information available on invertebrate fauna in the Victorian

mallee, my aim was to inform future work by providing a broad inventory of

Diptera and Hymenoptera in my study system, as these groups are the most

commonly recognised insect providers of ecosystem services, i.e. pollinators and

biological control agents. The focus of my work is on Diptera and Hymenoptera

that could provide critical pollination services to almond growers; however,

because of lack of knowledge on almond pollinators in Australia, I have sampled

the two orders generally and have included other families in my data that are not

commonly categorised as pollinators (e.g. parasitic or predatory flies and wasps),

as species from these families can pollinate flowers in some contexts (Armstrong

1979; Larson et al. 2001; Wäckers et al. 2005; Ssymank et al. 2008) and recent

evidence suggests that pollination and biological control services provided by

insects are synergistically linked (e.g. Bos et al. 2007; Lundin et al. 2013).

19

European honeybees (Apis mellifera) were removed from samples prior to

analysis because my focus was on non-honeybee potential pollinators. Evidence

from previous studies suggests that honeybees and non-honeybee pollinator

insects often show spatial complementarity in nesting and foraging locations,

especially when resources are abundant, as was the case in my system (Markwell

et al. 1993; Roubik & Wolda 2001; Paini et al. 2005; Brittain et al. 2013; Rollin et

al. 2013). Thus, the presence of honeybees in my system was unlikely to have had

a major impact on the abundance and richness of wild bees, wasps and flies found

at each site. After sorting insect samples into taxonomic orders, non-Apis bees

were identified to species or subgenus by a taxonomic expert (K. Walker,

Museum Victoria) and I identified wasp and fly individuals to family using taxon-

specific identification keys (Hamilton et al. 2006; Stevens et al. 2006). Family-

level analyses may provide coarser estimates of richness than species-level

approaches (Krell 2004; Lovell et al. 2007). However, given the largely

undescribed Australian insect fauna and the lack of taxonomic expertise or

knowledge on these groups (Austin et al. 2004; Raven & Yeates 2007; Batley &

Hogendoorn 2009; Cardoso et al. 2011), employing taxonomic surrogacy (i.e.

analysing orders, families, genera or morphospecies instead of species) is useful

in studies that aim to provide broad inventories or comparisons between

communities of insect fauna (Oliver & Beattie 1996; Pik et al. 1999; Obrist &

Duelli 2010).

20

21

PART II: ANALYSIS

22

23

Chapter 3: Almond orchards with living ground cover host more wild insect

pollinators

Abstract

Wild pollinators are becoming more valuable to global agriculture as the

commercial honeybee industry is increasingly affected by disease and other

stressors. Perennial tree crops are particularly reliant on insect pollination, and are

often pollen limited. Research on how different tree crop production systems

influence the richness and abundance of wild pollinators is, however, limited. I

investigated, for the first time, the richness and abundance of potential wild

pollinators in commercial temperate almond orchards in Australia, and compared

them to potential pollinator communities in proximate native vegetation. I

quantified ground cover variables at each site, and assessed the value of ground

cover on the richness and abundance of potential wild pollinators in commercial

almond systems focussing on three common taxa: bees, wasps and flies. More

insects were caught in orchards with living ground cover than in native vegetation

or orchards without ground cover, although overall species richness was highest in

native vegetation. Percent ground cover was positively associated with wasp

richness and abundance, and native bee richness, but flies showed no association

with ground cover. The strongest positive relationship was between native bee

abundance and the richness of ground cover plants. My results suggest that

maintaining living ground cover within commercial almond orchards could

provide habitat and resources for potential wild pollinators, particularly native

bees. These insects have the potential to provide a valuable ecosystem service to

pollinator-dependent crops such as almond.

*This chapter has been published as Saunders ME, Luck GW, Mayfield MM

(2013) Journal of Insect Conservation 17:1011-1025.

24

3.1 Introduction

The importance of plant diversity to insect pollinators has been discussed broadly

in the literature (e.g. Carreck & Williams 2002; Potts et al. 2003; Kremen et al.

2007; Hagan & Kraemer 2010; Ebeling et al. 2011; Pywell et al. 2011). In natural

environments, the spatial and temporal diversity of floral resources is often

positively correlated with the richness and abundance of wild insect pollinators

(Potts et al. 2003; Ebeling et al. 2008; Fründ et al. 2010; Grundel et al. 2010).

Evidence from agricultural systems also shows that insect pollinator abundance

increases in response to greater plant diversity in the landscape, such as proximate

remnants of native vegetation (see Ricketts et al. 2008; also Blanche et al. 2006;

Chacoff & Aizen 2006; Gámez-Virués et al. 2009; Carvalheiro et al. 2010;

Kennedy et al. 2013), managed intercropping (Altieri 2004; Thevathasan &

Gordon 2004; Letourneau et al. 2011), or management practices dedicated to

maintaining weeds/wildflowers in field margins (Carreck & Williams 2002;

Haaland et al. 2011).

Living ground cover is not common in conventional agricultural systems,

but is often an integral part of organic, multicropped or traditional farming

systems (Altieri 2004; Gomiero et al. 2011). The environmental benefits of

ground cover in agricultural ecosystems are numerous and include improving the

physical, chemical and biotic quality of the soil (Sharley et al. 2008; Grasswitz &

James 2009; Ramos et al. 2010), minimising erosion and excessive runoff (Gómez

et al. 2011; Kremer & Kussman 2011; Ruiz-Colmenero et al. 2011), and

increasing the number of beneficial insects and decreasing herbivorous pests

(Hooks & Johnson 2004; Gámez-Virués et al. 2009; Walton & Isaacs 2011). Yet

there are few studies linking wild insect pollinators in agricultural crop systems

with ground cover floral resource diversity, particularly in perennial tree crop

systems. Historically, studies of ground cover in orchard crops have focused on

natural enemies of crop pests rather than potential crop pollinators (e.g. Colloff et

al. 2003; Horton et al. 2003; Brown 2012).

Perennial tree crop systems are becoming more dominant in the global

agricultural landscape, with plantations of oil palm, rubber, coffee, cocoa and tree

nuts (e.g. almond, macadamia, cashew etc.) expanding rapidly, particularly

throughout tropical regions (Hartemink 2005) and in Australia specifically

25

(Australian Nut Industry Council n.d.). The area under commercial tree nut

cultivation in Australia was nearly 55,000 hectares in 2011 and future expansion

of this industry is likely, given current industry development efforts (Australian

Nut Industry Council n.d.). In particular, almond (Prunus dulcis Mill.) is

becoming increasingly important to Australia’s agricultural economy. The crop

contributes more to Australian nut production than the native macadamia and had

a farm-gate value of more than AU$300 million in 2012 (Almond Board of

Australia 2013).

Almond is one of the few crops that require 100% cross-pollination for

fruit production (Hill et al. 1985; Weinbaum 1985; Cunningham et al. 2002),

which means that managed European honeybee (Apis mellifera L.) hives have

become an essential part of production for commercial almond growers

(vanEngelsdorp et al. 2008). In addition, many of the other tree crops grown in

Australia on a commercial scale are reliant on pollinators for at least 50 % of their

production value (Cunningham et al. 2002). Yet there is a dearth of scientific

research into ecological and management issues associated with Australian tree

crop production, especially pollination ecology and the potential for wild

pollinators to benefit commercial systems. The reliance on honeybees in Australia

raises serious concerns for the future of pollinator-dependent crops, as impacts

such as the Varroa destructor mite, climate change, agricultural intensification,

and colony collapse disorder (CCD) increasingly affect the global honeybee

industry (vanEngelsdorp et al. 2008; Spivak et al. 2011; Henry et al. 2012) and are

likely to worsen in Australia in coming decades. Also, a recent meta-analysis has

shown that maximal fruit yields are not generally possible without the combined

services provided by managed honey bees and diverse wild pollinators (Garibaldi

et al. 2013).

Despite increased global interest in wild pollinators (Klein et al. 2012;

Garibaldi et al. 2013; Lebuhn et al. 2013) and the burgeoning research on

pollination as an essential ecosystem service (e.g. Klein et al. 2007; Kremen et al.

2007; Potts et al. 2010; Kennedy et al. 2013), current ecological knowledge of

almond pollination is dominated by research conducted in the United States,

particularly involving European honeybees (Degrandi-Hoffman 2001; Thomson

& Goodell 2001; Klein et al. 2012; Yong et al. 2012; Brittain et al. 2013). Non-

honeybee pollinators have proven successful in commercial almond orchards in

26

the United States, Japan and Spain (Bosch & Blas 1994), yet only a handful of

studies have investigated almond pollination in Australia (Hill et al. 1985;

Rattigan & Hill 1987; Vezvaei 1994; Vezvaei & Jackson 1995; Jackson 1996) and

these focus on the reproductive biology, gene flow or phenology of the flowers,

rather than non-Apis sources of pollination services. Australia has a unique and

endemic-rich insect fauna (Michener 1979; Austin et al. 2004; Batley &

Hogendoorn 2009), including many species important for the pollination of a

variety of crops (Heard 1999; Blanche et al. 2006; Gaffney et al. 2011; Lentini et

al. 2012). The dynamics of Australian almond systems are likely to differ

substantially from other major growing regions in California and Spain, because

of different local climates and ecological communities, and also from other semi-

arid locations where the almond is a native species (e.g. Israel; Mandelik & Roll

2009). Considering the economic value of the crop to Australia's economy, and

the possibility of disruption to managed pollination services if Varroa reaches

Australia, knowledge of potential wild pollinators in Australian almond

production systems is needed. Research into the links between management

activities and pollinator communities is particularly vital for the ecologically

sustainable management of almond crops, and will increase understanding of the

importance of wild pollinators to crop production worldwide (Potts et al. 2010;

Garibaldi et al. 2011; Mayer et al. 2011).

To my knowledge, this study is the first to sample wild pollinators in

commercial almond orchards in Australia, and one of the first to investigate the

relationship between wild pollinators and living ground cover in tree crop systems

anywhere in the world. Specifically, I asked the following questions: (1) How

does the richness and abundance of insect pollinators in commercial almond

orchards compare to wild pollinator communities present in proximate native

vegetation? (2) Is there a relationship between living ground cover and insect

pollinator richness and abundance within commercial almond orchards? I refer to

native bees, wasps and flies as potential pollinators (hereafter, pollinators).

Although I did not quantify pollination efficiency in this study, my target species

are the most recognised and frequent flower visitors and pollinators in a variety of

ecosystems (Kevan & Baker 1983; Kearns 2001; Larson et al. 2001) and I

observed flies and wasps visiting almond flowers during sampling.

27

3.2 Methods

In this paper, the following terms are used: ‘broadacre monoculture’ refers to a

cultivation system in which a crop is grown in a continuous, strict monoculture

(i.e. no other crop or managed plants within the physical boundaries of the

system) across an area of land greater than 100 hectares; ‘living ground cover’

refers to a carpet of living native and/or non-native herbs and grasses that is sown

or naturally-seeded, and is maintained within a cultivated orchard system.

3.2.1 Study area and system

This study was conducted in 2010 in northwest Victoria, Australia, in the

Victorian Murray Mallee Bioregion, specifically in the neighbouring Wemen—

Liparoo and the Colignan—Nangiloc districts (34° S 142° E). The study area is

semi-arid and is dominated by irrigated agriculture and areas of native mallee

vegetation on sandy, calcareous soils. During the sampling year, the region

recorded wetter than average rainfall – approximately 528 mm fell near Wemen—

Liparoo in 2010 (station 76000, long-term average = 314.9 mm p.a.), while

Colignan—Nangiloc received approximately 427 mm (station 76052, long-term

average = 290.4 mm p.a.). Average temperatures across both districts in the

months of sampling (August-September 2010) ranged from minimums of 5-7oC to

maximums of 15-20oC (Bureau of Meteorology 2013). Sampling sites were

located within five commercial almond orchards and proximate native mallee

vegetation, which is the dominant native vegetation type in this region (Figure

3.1). Orchards were managed as either broadacre monoculture orchards with bare-

ground floors (hereafter MONO) or plant-diverse orchards with living ground

cover throughout (hereafter GRASS) (Table 3.1). Native mallee sites were located

within patches of remnant mallee vegetation greater than 50 metres across and

within 500 m of an almond orchard (hereafter Native Liparoo (NL), Native

Wemen (NV) and Native Colignan (NC)). All mallee patches were either adjacent

to one of the sampled almond orchards or separated from an orchard by a

roadway, field headland or barren area. The three GRASS orchards were all

within 3 km of each other and the two MONO orchards were 7.5 km apart. All

orchards were located within the Sunraysia agricultural region, which is a mosaic

28

Table 3.1 Almond orchards and mallee sites used for sampling. ‘Mixed farm’ means that almond blocks were part of larger farms that cultivated multiple crops within their boundaries (e.g. citrus, other nuts or vegetables). Area, management, and matrix information are not given for Native sites, as sites were in native vegetation patches with varying locations. Ground cover and plant richness values are averages across all sampling sites within the relevant sampling location (see Methods for details of sampling). Sampling

location

Vegetation Type Area

(ha)

Management Surrounding

Landscape Matrix

Ground

cover

No. of

sampling

sites

% ground

cover

Absolute plant

richness

Liparoo Broadacre

monoculture

(MONO)

1002 Conventional (no

pesticide or

insecticide)

Native vegetation,

agricultural

None 15 0.007 (0.01) 0.3 (0.33)

Wemen Broadacre

monoculture

(MONO)

165 Conventional (no

pesticide or

insecticide)

Native vegetation,

agricultural

None 15 0.02 (0.03) 0.4 (0.43)

Colignan B Small, mixed farm

(GRASS)

76 Biodynamic Native vegetation,

agricultural

Complete 9 76.4 (11.01) 3.9 (0.65)

Nangiloc Small, mixed farm

(GRASS)

57 Biodynamic Agricultural Complete 8 92.0 (6.8) 3.6 (1.38)

29

Colignan C Small, mixed

farm

(GRASS)

160 Conventional

(low-intensity)

Agricultural Inter-row 10 63.4 (13.11) 3.0 (0.56)

Native

Liparoo (NL)

Native mallee

(Native)

- - - - 15 46.3 (43.4) 2.1 (1.84)

Native

Wemen (NV)

Native mallee

(Native)

- - - - 15 57.4 (22.8) 3.2 (1.14)

Native

Colignan

(NC)

Native mallee

(Native)

- - - - 13 60.1 (14.7) 3.0 (0.58)

30

landscape of irrigated crop systems and remnant areas of native mallee woodland,

black box (Eucalyptus largiflorens) woodland and river red gum (E.

camaldulensis) forests. The average proportion of natural vegetation within a 1

km radius around GRASS sites was 0.21 ± SD 0.08, and the average proportion

around MONO sites was 0.36 ± SD 0.17. The minimum distance between a

GRASS orchard and a MONO orchard was 35.1 km. Study locations were limited

by the availability of almond orchards maintaining ground cover – most

conventional almond growers maintain bare orchard floors, as this has historically

been advocated to prevent frost damage and competition for soil moisture (NSW

Agriculture, 2002; Wilkinson, 2012). A total of 30 MONO sites, 27 GRASS sites,

and 43 Native sites were successfully sampled during my study. All sampling

sites were chosen a priori, to ensure site independence, and all were at least 200

m apart. Sampling sites within orchards were spaced as broadly as possible so that

the entire area of each orchard was represented in the samples.

Figure 3.1 Representative photographs of the three vegetation types sampled in this study: (top left) broadacre monoculture almond orchards (MONO); (top right) plant-diverse orchards with permanent living ground cover (GRASS); and (bottom) native mallee vegetation (Native).

31

Liparoo and Wemen farms were owned and operated by a multinational company,

and were broadacre, conventional monoculture plantations that were

representative of commercial almond orchards common to the region. Herbicides

were applied (Glyphosate, Spray.Seed and occasionally Striker®) at label rates as

required by seasonal management regimes (Richmond, M. 2011, pers. comm., 10

June). No pesticides or insecticides were used. Bare compact soil was maintained

across the orchard floors in both these farms (Figure 3.1). The Colignan and

Nangiloc farms were owned and operated by a local family-run farm business.

Nangiloc and Colignan B were small, multifunctional biodynamic farms (a type of

holistic organic agriculture). Permanent living ground cover was consistent

throughout these orchards (Figure 3.1) and was managed by mowing when

necessary for access, usually before harvest (January-March). Colignan C was a

small, multifunctional conventional (low-intensity) orchard. Bare soil was

maintained along almond tree lines, comprising a strip approximately 1 m wide

surrounding the tree trunks, and permanent inter-row ground cover was managed

with mowing. No herbicides or pesticides were used in these orchards (Keens, I.

2010, pers. comm., 19 August). Mowing schedules were not recorded in this

study, but ground cover plants were well-established and flowered along tree lines

and throughout inter-row sections during sampling, as detailed below.

I did not include a biodynamic-conventional management comparison in

this study, as the sampled GRASS orchards were all small-scale farms operated

by the same family, with similar management strategies and in the same

geographical area, and the sampled MONO orchards were managed with very low

chemical usage. Some low-impact or small-scale farms operate under similar

principles and exhibit similar ecological characteristics to organic farms (e.g.

maintenance of in situ plant diversity; Francis & Porter 2011), but cannot be

labelled as ‘organic’ due to policy and certification systems. Hence, I focus my

investigation on the relationships between ecological characteristics in the crop

systems (Letourneau & Bothwell 2008; Winqvist et al. 2012), rather than the

anthropocentric label of ‘management type’.

32

3.2.2 Insect Sampling

Insects were sampled during the almond flowering period (August-September)

over 10 days in 2010. Almond flowers for a very short time (typically 2-3 weeks

per tree) in response to weather cues (Hill et al. 1985), restricting the most

suitable sampling time. Each sampling site was sampled once for approximately

8-10 daylight hours (depending on seasonal conditions) using a set of three pan

traps – one blue, one yellow and one white bowl. Traps were made from plastic

picnic bowls (Woolworths Home Brand, Bella Vista NSW) painted with UV-

bright fluorescent blue or yellow paint (White Knight Paints, Lansvale NSW) or

left white, and filled with 200 mL of water and a drop of dishwashing detergent to

break surface tension. Because of the cool winter temperatures during sampling, I

set traps out soon after sunrise (8-10 am) and collected them just before nightfall

(3.30-6 pm) in the order they were set, to capture maximum insect activity during

the warmest part of the day. All sampling days were fine, sunny or partly cloudy,

with light or no breeze. Pan trapping may not provide full representation of the

sampled pollinator community, as the effectiveness of the traps can be influenced

by trap placement, surrounding habitat structure, proximate resource availability

or even insect colour vision (Cane et al. 2000; Baum & Wallen 2011; Saunders &

Luck 2013; see Appendix 1). However, the method can be more useful than other

insect collection methods, such as on-site visual identification, as it does not

suffer from ‘collector bias’ (Westphal et al. 2008) and is efficient for establishing

baseline data for previously unstudied ecosystems. To increase the likelihood that

pan trap data are ecologically meaningful, it is important to ensure that bowl

colour and field placement are consistent and relevant to the context of the study

and research questions (Tuell & Isaacs 2009; Vrdoljak & Samways 2012;

Saunders & Luck 2013). I chose white, yellow and blue traps, as white was the

predominant flower colour in orchards and yellow was the most common flower

colour in local mallee woodlands. I also included blue traps, because this colour

has been recommended for attracting bees in some environments, but has not been

widely used to collect native bees in Australia (see Appendix 1). At each site, a

set of traps (one of each colour) was placed in a triangular array on open ground 1

m from a flowering tree, to provide maximum visibility.

33

Upon collection of the traps, insects were removed from the water and

placed into 70% ethanol, before being transported to the laboratory. I pooled

catches from each set of traps (blue, yellow, white) to count individuals per site.

Native bees were identified to species by a taxonomic expert (K. Walker, Museum

Victoria) and Iidentified wasps and flies to family level using specialised

Hymenoptera and Diptera identification keys (Hamilton et al. 2006; Stevens et al.

2007). Australia's invertebrate fauna is one of the most under-described in the

world and many insect orders require critical taxonomic revision, including the

Diptera (Austin et al. 2004; Raven & Yeates 2007). Family-level identification is

an appropriate taxonomic surrogate when research goals include comparing broad

estimates of ecosystem service providers (Oliver & Beattie 1996; Obrist & Duelli

2010), and I considered this a suitable level of identification for this study, given

the challenges of identifying Australian wasp and fly taxa to the species level and

the uncertainty about the taxonomy of these groups.

3.2.3 Ground cover sampling

Compositional differences in herbaceous plant species were not quantified in this

study; however, many of the exotic (introduced to Australia) ground cover species

within GRASS orchards were flowering during the study (e.g. Persian speedwell

(Veronica persica), Capeweed (Arctotheca calendula), dog violet (Viola spp.),

Oxalis spp.). In MONO orchards, the few ground cover species present were non-

flowering seedlings (e.g. self-propagated almonds, stinging nettle (Urtica dioica),

wild lettuce (Lactuca spp.) and radish (Raphanus spp.)). At native mallee sites,

most available flowers were in middle- or upper-storey trees and shrubs.

Percentage ground cover and plant species richness were quantified at each

sampling site once on the same day that insects were collected. At each site, I

used four 0.5 metre x 0.5 metre quadrats placed randomly 1-5 m from the centre

of each set of pan traps, in each cardinal direction (N, S, E, W). Percentage

ground cover (hereafter ground cover) and absolute plant species richness

(hereafter plant richness) values from the four quadrats (N, S, E, W) in each site

were averaged to obtain mean percentage ground cover and mean plant richness at

each site (Table 1).

34

3.2.4 Data analysis

Analyses were conducted on combined Hymenoptera and Diptera samples

(response variable: Total Insects), and each separate pollinator group (response

variables: flies; native bees; wasps). I used PAST 2.17 (Hammer et al. 2001) for

statistical analyses, unless indicated otherwise. Abundance data were

overdispersed ecological count data with a non-normal distribution. Hence, I used

nonparametric analysis where applicable rather than transforming the data to suit

parametric analyses, as this approach is increasingly recommended (McArdle &

Anderson 2004; O’Hara & Kotze 2010; Warton & Hui 2011). For each vegetation

type, I calculated the interquartile range for each insect group’s abundance (as a

measure of statistical dispersion) and the number of taxa per group found in each

vegetation type, as a proportion of the total number of taxa of each group

collected across the study area. I tested for differences in insect abundance (total

counts of individuals) between vegetation types with Mann-Whitney (U) pairwise

comparisons. To see how wild pollinator communities in native vegetation

compared with communities in almond orchards, I calculated Whittaker’s measure

of shared richness (βw) between vegetation types, where 0 means site pairs have

the same species composition and 1 means site pairs have completely different

communities. Because sampling effort was limited and sample sizes differed

between vegetation types, I computed extrapolated species accumulation curves

(EstimateS version 9.1; Colwell 2013) to compare estimated total richness of each

insect group between vegetation types (Colwell & Coddington 1994; Gotelli &

Colwell 2001; Colwell et al. 2004). Native sites had the largest reference sample

(n = 43), so richness estimates for GRASS and MONO sites were extrapolated to

a sample size of 43, which is less than the maximum recommended extrapolation

limit (Melo et al. 2003; Colwell et al. 2012). Extrapolation curves were calculated

using the Chao1 target richness estimator (Colwell et al. 2012). Using these

estimates, I plotted estimated total family richness for flies and wasps and

estimated total species richness for native bees (based on 100 randomisations) by

vegetation type, as a function of the number of sampled individuals. This

approach is suitable for analysing any taxonomic level and is recommended for

sample-based abundance data, and for evaluating theoretical predictions or models

(Gotelli & Colwell 2001; Colwell et al. 2012).

35

In addition to estimating pollinator richness, I also created diversity

profiles for pollinator groups in each vegetation type using PAST 2.17 (Hammer

et al. 2001). Flies and wasps were analysed as family groups, and bee species

were combined into genera, as there is the potential for taxonomic similarity at the

species-level to affect the results of diversity profiles (Leinster & Cobbold 2012).

Diversity ordering (or 'profiling') measures and compares taxonomic richness

between communities using multidimensional families of diversity indices rather

than a single index (Patil & Taillie 1982; Tóthmérész 1995; Liu et al. 2007;

Leinster & Cobbold 2012). Calculations based on one index can show much less

complexity than is truly present in a study system (Magurran 2004; Gattone & Di

Battista 2009), which can result in inconsistencies in analyses and interpretations.

Diversity ordering reduces this effect and provides a more realistic comparison of

community diversity by measuring a series of related indices and constructing a

diversity ‘profile’ of the community (Tóthmérész 1995; Liu et al. 2007; Leinster

& Cobbold 2012). I used the exponential of Rényi's diversity ordering index,

which plots the community’s diversity order against a continuous sensitivity

parameter (α) (Tóthmérész 1995; Hammer et al. 2001). The parameter is based on

true diversities (Hill numbers) and represents how sensitive the diversity order is

to the number of species (Jost 2006). The left side of the graph (α = 0) represents

the absolute richness of the community and the number of rare species, while the

right side of the graph (α = 4) represents dominance and the number of common

species (Hill 1973; Leinster & Cobbold 2012). Where a sample's profile sits

completely above any other plotted lines, the higher sample is unequivocally more

diverse than other samples. Where two lines intersect, the relevant samples cannot

be explicitly separated from one another in terms of overall diversity. However,

the intersection point of the two lines, when interpreted relative to α, shows how

the communities differ in terms of dominance or richness (Leinster & Cobbold

2012).

To explore the overall relationship between ground cover in almond

orchards and abundance and richness of each insect group, I conducted

generalised linear modelling in IBM SPSS 20 (IBM Corporation 2011). For

pollinator richness, I calculated Menhinick’s diversity index for fly and wasp

families and bee species at each site, as this index takes individuals per

species/family into account and is suitable for unequal sample sizes (Menhinick

36

1964). For abundance data, I used negative binomial regression due to the nature

of the overdispersed count data (Seavy et al. 2005; Sileshi 2006; O’Hara & Kotze

2010). Richness data were analysed with linear models, as these data were

normally distributed. The maximum number of parameters I could include in any

model (to avoid over-fitting) was five, because of the small sample size:

n[GRASS+MONO] = 57 (Burnham & Anderson 2002). Based on my research

questions, the focal predictors were ‘ground cover’ and ‘plant richness’; however,

as these two variables were strongly correlated (r = 0.90, p < 0.0001) they were

not included in the same model. Orchard sites were distributed at different

distances to natural vegetation, so I also included ‘distance to natural vegetation’

as a predictor to determine if landscape context influenced pollinator responses to

ground cover. However, I do not consider landscape context further, as the goal

was to identify the extent of the relationship between potential pollinators and

ground cover vegetation on orchard floors in my study system. I computed models

for each predictor alone, an additive and an interaction model for each ground

cover variable with the landscape variable, and an intercept-only model for a

baseline comparison, and used an information-theoretic approach (Burnham &

Anderson 2002) to identify the best explanatory model for each insect group

based on the research questions. For richness models, I calculated the second-

order Akaike’s Information Criterion for small samples (AICc) and for abundance

models, I modified this criterion for overdispersed count data (QAICc); hereafter,

both IC values are referred to as AICc. I also calculated Akaike weights (wi) for

each model, which represent the probability that the model is the best model for

the data. Within each model set, I compared differences between AICc values (Δ

AICc) to determine the explanatory model that provided the best fit for each insect

data set and any other models with substantial support (Δ AICc ≤ 2; Burnham &

Anderson 2002).

To determine how pollinator assemblages compared within vegetation

types, I calculated the following similarity/distance matrices, based on the 28

pairs of study locations (individual orchard or mallee patch; e.g. Wemen/Colignan

B; Wemen/Nangiloc; Wemen/NC etc.): (i) Bray-Curtis (SBC) similarities for

pollinator abundance data sets (bee species, fly and wasp families) (Faith et al.

1987; Clarke & Ainsworth 1993); and (ii) Euclidean distances for average percent

ground cover and average plant richness (Clarke & Ainsworth 1993). I tested for

37

correlations between these indices with a Spearman's rank test, to determine if

between-site similarity in ground cover was related to similarity of pollinator

communities (Clarke & Ainsworth 1993). I also used hierarchical cluster analysis

(Bray-Curtis similarity and unweighted paired-group linkages) to create

dendrograms comparing taxonomic similarity between bee, fly and wasp

assemblages at each study area.

3.3 Results

3.3.1 Comparison of wild potential pollinator communities between vegetation

types

Across all sites, I collected 3251 fly individuals from 25 families, 144 wasp

individuals from 19 families, and 90 native bee individuals from 12 species.

Lasioglossum (Chilalictus) globosum (Smith, 1853) was the most common bee

species overall (25 individuals) and was collected in both GRASS orchards and

mallee vegetation. No native bees were collected in MONO orchards. Tachinidae

and Calliphoridae were the most abundant fly families, and were collected in all

vegetation types (Appendix 2 Table A2.2). The most abundant wasp families

(Diapriidae and Braconidae) were also collected in all vegetation types, while six

bee species and five wasp families collected in mallee vegetation were not present

in almond orchards (Appendix 2 Table A2.2). The interquartile range for each

pollinator insect group was highest in GRASS orchards: Flies – GRASS (104)

MONO (9.75) Native (10.88); Wasps – GRASS (3) MONO (1) Native (1); Native

bees – GRASS (2) Native (1). Based on the total number of taxa (family or

species) collected across the study area, Native sites had the highest proportion of

recorded taxa for each insect group, but the number of fly families at GRASS sites

was equal to Native sites: Flies – GRASS (0.80) MONO (0.64) Native (0.80);

Wasps – GRASS (0.63) MONO (0.37) Native (0.79); Native bees – GRASS

(0.50) Native (0.92).

On average, GRASS sites had higher total pollinator abundance (mean

92.3 ± 82.2) than either Native (mean 14.4 ± 12.2; p = 0.02) or MONO sites

(mean 12.5 ± 8.6; p < 0.001). Average abundances for all insect groups were

significantly higher in GRASS sites than in Native or MONO sites (Table 3.2).

38

Pairwise comparisons showed that the abundance of Total Insects in mallee

vegetation was most similar to MONO orchards (U = 620.5, z = -0.27, P = 0.788),

as was wasp (U = 587, z = -7.03, P = 0.482) and fly abundance (U = 638, z = -

0.07, p = 0.94). Native bee abundance was significantly different between all three

vegetation types (GRASS ≠ MONO U = 165, z = -4.84, P < 0.0001; GRASS ≠

Native U = 408, z = -2.33, P = 0.019; MONO ≠ Native U = 450, z = -3.27, P =

0.001). Wasp communities at Native sites were most similar to GRASS sites (βw =

0.33) and least similar to MONO sites (βw = 0.45). Wasp community similarity

between GRASS and MONO sites was βw = 0.37. Fly communities at Native sites

were more similar to MONO sites (βw = 0.22) than GRASS sites (βw = 0.25), and

very similar between GRASS and MONO sites (βw = 0.17). Native bee

community similarity between GRASS and Native sites was βw = 0.41.

Table 3.2 Average abundance and richness of potential wild pollinator insects collected per site, in native mallee vegetation (Native), broadacre monoculture almond plantations (MONO), and plant-diverse almond orchards (GRASS) in the Murray Mallee region of north-west Victoria. Richness values are based on the Menhinick diversity index (these were not used to compare richness between vegetation types, but are given here for reference). Asterisks denote where abundance was significantly higher than the other two vegetation types.

Abundance, mean (± SD) Richness, mean (± SD)

Fly Wasp Native bee Fly Wasp Native bee

Native 12.8 (11.81) 0.9 (1.74) 0.7 (1.2) 1.4 (0.5) 0.5 (0.7) 0.4 (0.6)

MONO 11.5 (8.5) 1.0 (1.35) 0 1.4 (0.6) 0.6 (0.6) 0

GRASS 87.2 (83.3)* 2.8 (2.1)* 2.3 (4.2)* 1.3 (0.6) 1.1 (0.8) 0.7 (0.6)

* p < 0.05

Comparison of extrapolated richness curves suggested that my sampling effort

may not have captured all taxa present in my study system, as only the curves for

native bees at GRASS sites and flies at Native sites approached an asymptote.

Native mallee sites had more unique bee species and fly and wasp families than

either orchard type, and GRASS sites had consistently higher abundance across all

insect groups (Figure 3.2). There was no significant differences in wasp richness

between vegetation types (Figure 3.2a), but fly and bee richness were significantly

higher at Native than at GRASS sites (Figure 3.2b-c).

39

Figure 3.2 Accumulation curves showing the number of (a) wasp families (b) fly families and (c) native bee species per number of individuals collected in each vegetation type. Broken lines indicate the 95% confidence intervals for each curve.

0

5

10

15

20

25

0 20 40 60 80 100 120

No.

of w

asp

fam

ilies

(a)

Native

GRASS

MONO

0

5

10

15

20

25

30

35

0 500 1000 1500 2000 2500 3000 3500 4000

No.

of f

ly fa

mili

es

(b)

GRASS

MONO

Native

0

2

4

6

8

10

12

14

0 20 40 60 80 100 120

No.

of n

ativ

e be

e sp

ecie

s

No. of collected individuals

(c)

Native

GRASS

40

Community diversity profiles were mostly consistent with these patterns

(Figure 3.3). Wasp and bee communities were more diverse in native mallee

vegetation than in almond orchards (Figure 3.3a,c). Fly diversity profiles from all

three vegetation types overlapped, but Native sites had slightly higher richness of

rare families (Figure 3.3b). Between orchard types, wasp and fly diversity were

similar, but GRASS orchards had higher richness (more rare families) while

MONO orchards had slightly higher dominance of common families (Figure 3.3a-

b).

3.3.2 Insect pollinator and ground cover association

The predictor ‘distance to natural vegetation’ was present in highly-ranked models

for fly abundance, wasp abundance and richness, and native bee richness, but

there was no evidence that this predictor had a strong influence over pollinator

relationships with ground cover in almond orchards (Table 3.3). Fly richness and

abundance were not associated with either percent ground cover or plant richness.

Percent ground cover was positively associated with bee and wasp richness and

wasp abundance. Native bee abundance had a strong positive association with

plant richness, and this predictor also had some influence on variation in wasp

richness (Table 3.3).

Taxonomic similarity between bee communities increased with similarity

in percent ground cover (rs = 0.67, P < 0.001) and plant richness (rs = 0.71, P <

0.001) between sites. Between-site correlations of differences in ground cover

with wasp and fly community similarity were weak (rs < 0.3) and not significant

(p > 0.1). Wasp communities in the monoculture Wemen orchard were more

similar to GRASS orchards (SBC: Wemen/Nangiloc = 0.74; Wemen/Colignan B =

0.51; Wemen/Colignan C = 0.59), than to the other MONO orchard, Liparoo (SBC

= 0.41). Fly communities at the two biodynamic orchards (Colignan B and

Nangiloc) were more similar to each other (SBC = 0.68) than to communities at

any other location (average between-site SBC = 0.25). Bee communities at

Colignan C were significantly different to bee communities at any other location

(maximum between-site SBC = 0.18) (see Figure 3.4).

41

Figure 3.3 Diversity profiles for (a) wasp families (b) fly families and (c) native bee genera collected in each vegetation type. The left hand side of the plotted line (α = 0) represents absolute richness and relative number of rare species in the community and the right hand side (α = 4) represents dominance and relative number of common species.

0

2

4

6

8

10

12

14

Was

p di

vers

ity

(a)Native

GRASS

MONO

0

5

10

15

20

25

Fly

dive

rsity

(b)

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

0 1 2 3 4

Nat

ive

bee

dive

rsity

Sensitivity paramater (α)Richness Dominance

(c)

42

Table 3.3 Parameter estimates with confidence intervals from negative binomial regression (abundance) and linear regression (richness) analysis testing the relationship between insect groups and ground cover variables. R2 coefficients are given for linear models only. The highest-ranked models (Δ AICc ≤ 2) are presented for each insect response variable. Where the ‘constant only’ model was the highest-ranked model, only the second-highest ranked model has been given for comparison. Akaike weights (wi) for each model indicate the probability that the given model is the best fit for the data. Insect response Model AICc Δ AICc wiFly Abundance Constant only (-0.11, -0.14 to -0.09)

D (0, 0 to 0)

325.5

327.1

0

1.5

0.37

0.17

Richness GC (-0.003, -0.007 to 0.001); R2 = 0.04

Constant only (1.33, 1.18 to 1.49)

107.9

108.1

0

0.2

0.25

0.23

Wasp Abundance GC (0.006, 0.003 to 0.01)

GC (0.007, 0.003 to 0.01) + D (0.001, 0 to 0.002)

201.9

202.3

0

0.4

0.43

0.34

Richness GC (0.006, 0.001 to 0.01); R2 = 0.10

GC (0.006, 0.002 to 0.01) + D (0.001, 0 to 0.002); R2 = 0.13

PR (0.12, 0.008 to 0.22); R2 = 0.07

127.9

128.3

129.7

0

0.4

1.8

0.31

0.26

0.13

Native bee Abundance PR (1.18, 0.52 to 1.8) 88.0 0 0.60

Richness GC (0.007, 0.004 to 0.01); R2 = 0.31

GC (0.007, 0.005 to 0.01) + D (0, 0 to 0.001); R2 = 0.32

72.2

73.6

0

1.4

0.54

0.27

GC = percent ground cover; PR = plant richness; D = distance to natural or semi-natural vegetation. All models include the constant as a parameter.

43

Figure 3.4 Bray-Curtis similarities between (left) wasp families (middle) fly families and (right) native bee species per site group. Native sites: NL, NV, NC; MONO sites: WEM, LIP; GRASS sites: CC, COL, NAN.

44

3.4 Discussion

I found that overall richness of all three potential pollinator groups was highest in

native mallee vegetation. However, almond orchards with permanent living

ground cover throughout the orchard supported greater wild pollinator richness

and abundance than orchards without ground cover. Total insect abundance, wasp

abundance and fly richness and abundance at mallee vegetation sites were more

similar to MONO orchards than GRASS orchards, but no native bees were caught

in MONO orchards during the sampling period. Mallee wasp families were more

similar to wasp families in GRASS orchards than to those in MONO orchards.

Flies were not influenced by ground cover in almond orchards, but ground cover

did have a positive association with wasp and native bee abundance and richness.

The strongest positive relationship was between native bee abundance and

richness of herbaceous ground cover plants. Native bee community similarity

between sites was strongly correlated with similarity in ground cover attributes,

but wasp and fly community similarity were not correlated with ground cover

similarity between sites. Cluster analysis (Figure 3.4) showed variation in

pollinator assemblages within vegetation types. Wasp assemblages at the

monoculture Wemen orchard were more similar to GRASS orchards than the

other monoculture orchard, while bee abundance in one GRASS orchard was

higher than in any other study area. In addition, no native bees were caught at any

MONO site, despite bees being present in proximate native vegetation. I tested for

the influence of broader landscape factors on insect response but this did not

override pollinator relationships with ground cover in my system. However, the

uncertainty of some pollinator—ground cover relationships, the lack of

association between ground cover and fly communities, and the pollinator

community differences between orchards, indicate that other factors may be

influencing some pollinator groups in this system.

Wild insect pollinators often prefer open, heterogeneous, ‘moderately’

anthropogenic habitats, such as grasslands, fallow agricultural fields, meadows or

cultivated gardens over more dense forests (Hagan & Kraemer 2010; Samnegård

et al. 2011; Winfree et al. 2011). In particular, most solitary bee species require

access to multiple heterogeneous habitats throughout the year (Mandelik et al.

2012), as they prefer herbaceous floral diversity for food sources and fallen wood,

45

bare soil or other structures for nesting (Hoehn et al. 2010; Gruber et al. 2011).

Most modern conventional agricultural systems do not offer this heterogeneity of

food and nesting resources, providing only one species of flower with a short

blooming season, homogeneous soil conditions and trees pruned of any dead

wood that may provide nesting substrate. This is especially true for perennial tree

crop systems, such as fruit and nut orchards, when they are managed intensively

as strict monocultures with no ground cover or intercropped vegetation. Despite

intensive management, orchard systems are still structurally similar to some types

of dense forests (such as temperate deciduous forests), which also generally

harbour low bee abundance (Winfree et al. 2007; Hoehn et al. 2010; Proctor et al.

2012). Unlike orchard crops in North America and Europe, which are often

structurally similar to native deciduous forest systems, Australian almond

orchards are very different to surrounding evergreen mallee woodlands. These

differences could potentially influence the extent of pollinator services that may

come from insects living in mallee and are worth exploring further in the future.

The monoculture orchards I sampled had bare orchard floors, which could

provide suitable nesting substrate for some wasp and bee species, but had no floral

resource diversity. In contrast, the biodynamic orchards had a carpet of

floral/herbaceous diversity, but few areas of bare ground. Only one GRASS

orchard, Colignan C, had inter-row herbaceous ground cover and bare ground

along the tree line. Interestingly, I caught more wasps and native bees in Colignan

C than at the other GRASS sites, even though Colignan C had less consistent

ground cover than the other plant-diverse orchards. Pan traps were more visible at

this orchard than in the other two GRASS orchards, which could have increased

catch rates (Laubertie et al. 2006). However, it is unlikely that my results were

unduly influenced by trap visibility, as traps were more visible in MONO

orchards where ground cover and understorey vegetation were sparse and overall

heterogeneity was very low, yet I caught no bees at any MONO site. These results

more likely indicate that herbaceous ground cover can support many wild insect

pollinators in commercial orchards and that a diverse array of within-orchard

habitat elements may interact with the presence of ground cover to sustain insect

communities in the long-term; however, more sites with similar attributes would

need to be sampled to validate this possibility.

46

Although living ground cover in Australian almond orchards has

historically been considered to have a negative influence on moisture availability,

orchard microclimates and harvest efficiency (Wilkinson 2012), finding a balance

between maintaining understorey vegetation and management efficiency could

have positive implications for overall ecosystem function and productivity.

Ground cover in almond orchards can improve soil quality and biological activity

(Ramos et al. 2010), which is particularly important in arid ecosystems (Whitford

1996), and it can also support important natural control agents for almond pests

like the navel orange worm (Eilers & Klein 2009). There has been very little

research into direct links between permanent living ground cover in commercial

orchards and pollinator richness and abundance; however, my results correlate

with current knowledge of similar study systems. The only directly comparative

study I am aware of found greater wild bee abundance in a weedy section of an

almond orchard in Israel compared to a non-weedy part of the orchard, although

the authors do not provide further details on how vegetation composition differed

between the two sections of orchard (Mandelik & Roll 2009). Studies from other

fruit or nut orchard systems have also shown that ‘natural enemy’ insect

abundance (including wasps) increased as within-orchard plant diversity (cover

crops or weedy ground cover) was enhanced or left to grow unmanaged (Tedders

1983; Horton et al. 2003; Brown 2012). Evidence from European agroecosystems

showed that wild bee abundance in apple orchards responded positively to

increased extent of fallow areas (unmanaged, grassy or weedy) surrounding

orchards (Gruber et al. 2011) and that wasps were more abundant in abandoned

(weedy) orchards than in managed orchards (Steffan-Dewenter & Leschke 2003).

Crop pollinator abundance and diversity within vegetable or seed crop systems

also responds positively to non-crop plant diversity within the field area (e.g.

Jones & Gillett 2005; Gemmill-Herren & Ochieng 2008; Carvalheiro et al. 2011).

Many studies have shown the positive effect of nearby native vegetation

on pollinator richness or abundance in agricultural systems (e.g. Ricketts et al.

2008; Watson et al. 2011; Lentini et al. 2012; Lowenstein et al. 2012), including

in North American almond orchards (Klein et al. 2012). However, very few of

these studies provide comparative analysis of pollinator abundance in native

vegetation relative to abundances within the crop system. A notable exception is

Kehinde & Samways’ (2012) recent study from South Africa, which investigated

47

insect pollinators in organic and conventional vineyards as well as proximate

natural vegetation. They showed higher pollinator abundance in both organic

farms and natural vegetation than in conventional farms. Although their analysis

focussed on the ‘organic vs. conventional’ responses of wild pollinators, the

authors note that organic vineyards contained flowering cover crops, which were

not present in conventional vineyards, and they suggest this may have been the

cause of higher insect abundance in organic vineyards (Kehinde & Samways

2012). It is clear that plant-diverse crop systems harbour higher pollinator

abundance, yet, historically, relevant studies have focused on the correlation of

plant and insect diversity isolated within the crop system itself. Some authors

have hypothesised that ‘weedy’ crop systems (compared to intensive or

monoculture systems) may have enormous conservation potential for local insect

pollinators (Gemmill-Herren & Ochieng 2008; Hyvönen & Huusela-Veistola

2008; Pinke & Pal 2009), but there is scope for further detailed investigation into

how plant-diverse crop systems actually compare to local native vegetation as

suitable pollinator habitat.

In conclusion, maintaining permanent living ground cover as part of

commercial orchard management practices has the potential to support higher

wild pollinator abundance, particularly native bees, which has important

conservation implications and could also supplement pollination services for

pollinator-dependent almond crops. The significance of my findings would be

augmented by further investigation of pollinator visitation rates and pollen

transfer potential. In addition, further research in a variety of orchard systems

would determine whether certain insect pollinator species benefit more from

living ground cover than others. It is also important to consider that a suite of

interrelated environmental variables may be of combined benefit to pollinator

communities, rather than one single variable on its own. There is abundant scope

for further research, in both my study system and other agroecosystems, linking

wild insect pollinator communities with combinations of related environmental

variables (e.g. ground cover presence/diversity and habitat structure) in

agroecosystems. Nonetheless, this study has shown that there is substantial

conservation potential for orchards with permanent ground cover, as unmanaged

potential insect pollinators, especially native bees, were more abundant in plant-

diverse orchards than in native mallee vegetation. Long-term studies in multiple

48

types of plant-diverse orchard systems will be useful for determining whether this

trend could motivate valuable pollinator conservation initiatives within the

orchard industry.

49

Chapter 4: Relationships between wild pollinators in flowering almond

orchards and proximate unmanaged vegetation

Abstract

Resource heterogeneity and connectivity is critical for maintaining diverse

pollinator communities in landscapes dominated by tree crop plantations, which

create large areas of spatial and temporal homogeneity. I collected potential wild

pollinators (native bees, wasps, and flies) in flowering almond orchards within a

heterogeneous (Complex) and homogeneous (Simple) agricultural landscape in

southern Australia. I examined relationships between wild pollinators and

vegetation types immediately surrounding each site, particularly natural

vegetation that provided permanent floral resources. Abundance and richness of

most individual pollinator groups were negatively associated with the proportion

of natural vegetation near sites in the Complex landscape and positively

associated with natural vegetation near sites in the Simple landscape, although the

strength and significance of the relationships varied. In the Simple landscape,

most individuals of the key pollinator taxa (native bees and hoverflies) were

collected in patches of natural vegetation inside the almond plantation, rather than

within blocks of almond trees. These results suggest that wild pollinators visiting

crops in homogeneous landscapes are more dependent on proximity to permanent,

unmanaged resources than wild pollinators using heterogeneous landscapes. More

research is needed into how landscapes dominated by tree crop monocultures

influence biodiversity and the provision of essential ecosystem services.

50

4.1 Introduction

Large-scale intensification of agricultural landscapes severely impacts ecological

processes and farmland biodiversity (Matson et al. 1997; Tilman 1999) and can

have widespread effects on community composition (e.g. Ekroos et al. 2010) and

the provision of ecosystem services (e.g. Marino & Landis 1996). To address

these effects, agricultural landscapes can be managed to increase vegetation

complexity and diversity, which can enhance biodiversity levels and ecosystem

function (Benton et al. 2003; Tscharntke et al. 2005). Vegetation heterogeneity is

particularly important for wild insect pollinators in agricultural landscapes and,

although some pollinators may be attracted to the temporary surplus of flowers in

crop fields (Holzschuh et al. 2013), the long-term establishment of a diverse

pollinator community depends on multiple foraging and nesting sites being

available throughout the year (Kremen et al. 2007). Numerous studies have

examined relationships between pollinator communities and the amount of

particular habitat types within agricultural mosaics, such as semi-natural

grasslands (Öckinger & Smith 2007), natural forest (Garibaldi et al. 2011 and

references therein) or sown wildflower strips around crop fields (Haaland et al.

2010 and references therein) and retention of unfarmed habitat is considered

essential for maintaining diverse pollinator communities around farmland.

However, a matrix that provides multiple types of habitat with different, but

complementary, levels of spatial and/or temporal heterogeneity is often more

important for the long-term establishment of wildlife communities than a single

permanent habitat type (Westrich 1996; Williams & Kremen 2007; Vasseur et al.

2013).

This type of matrix, providing resource heterogeneity and functional

connectivity, is particularly important in landscapes dominated by perennial tree

crop plantations, such as orchard fruits, because most fruit varieties depend on

insect pollination to set fruit (Kevan 1999); therefore, the presence of a diverse

pollinator community during crop flowering can be essential for some growers

(Garibaldi et al. 2013). Yet, unlike arable crop mosaics under short rotation that

can provide temporal heterogeneity in the absence of spatial variation (Vasseur et

al. 2013), tree crop plantations are generally embedded in the landscape for more

than 15 years, creating a comparatively ecologically stable, semi-permanent

51

ecosystem (Tedders 1983; Altieri 1999). If plantations are managed as intensive

monocultures, they create areas of permanent spatial and temporal homogeneity in

the landscape that are inhospitable to pollinators for most of the year (Sheffield et

al. 2008; Marini et al. 2012) and could potentially influence the structure of

invertebrate communities in the long-term (Harding et al. 1998; Tscharntke et al.

2002). Although the conservation value of tree crop agriculture has long been

recognised (Smith 1916; Altieri et al. 1983; Ryszkowski & Kędziora 1987), the

ecological role of tree crop plantations in agricultural landscapes has received less

attention in the scientific literature than annual crops (Hartemink 2005).

Historically, studies of orchard crop systems have focused on within-orchard

management practices to support beneficial ecosystem services like biological

pest control (e.g. Bugg & Waddington 1994; Brown 1999; Horton et al. 2003). To

inform biodiversity conservation and ecological management goals, there is a

need for more studies that investigate how established orchard crop plantations

influence communities of ecosystem service providers through interactions with

the surrounding landscape and ecosystems, particularly local pollinator

communities present during the tree crop’s brief flowering period (sensu Watson

et al. 2011; Marini et al. 2013).

Here, I focus on plantations of almond (Prunus dulcis Mill.) trees in a

semi-arid region of southern Australia. Almond production contributes to more

than half of Australia’s annual nut production and the country is the second-

largest almond producer in the world, with broadacre plantations expanding

across Australia’s major agricultural regions in the last few decades (Wilkinson

2012; Almond Board of Australia 2013). Cultivated almond trees flower

synchronously for 3-4 weeks at the end of winter (Hill et al. 1985) and are totally

dependent on insect pollination to set fruit, yet there is little published information

on wild pollinators of almond flowers in Australia or how communities of these

insects are influenced by almond plantation management (but see Saunders et al.

2013). Over 60 % of Australian commercial plantings are established in the semi-

arid mallee region of northwest Victoria (Almond Board of Australia 2013), an

area of high conservation value (Mallee Catchment Management Authority 2008).

Semi-arid landscapes are naturally dynamic and heterogeneous (Sala &

Lauenroth 1982; Aguiar & Sala 1999), especially in Australia (Ludwig &

Tongway 1995; Morton et al. 1995) where they are being increasingly impacted

52

by agricultural intensification. Yet, despite much discussion of how agricultural

fragmentation impacts natural ecosystems in other parts of Australia (e.g.

Lindenmayer & Hobbs 2004; Haslem & Bennett 2008; Smith et al. 2013), there is

very little published research into how wildlife communities in dynamic,

heterogeneous semi-arid ecosystems like the mallee are influenced by large areas

of homogeneous crop plantations (e.g. Driscoll & Weir 2005; Luck et al. 2013).

This study aims to contribute to some of these knowledge gaps and is part of the

first investigation of potential wild pollinators available to almond crops in

Australia’s main almond growing region, the Victorian Murray Mallee. In this

chapter, I examine whether the abundance and richness of potential wild

pollinators found in almond plantations during flowering are associated with

different vegetation types in immediate proximity to sites. Due to the lack of

ecological information available on my study system, I sampled flowering

orchards in two different types of agricultural landscape (heterogeneous and

homogeneous) and concentrated on documenting patterns of association between

potential wild pollinators collected in flowering orchards and the presence of

unmanaged vegetation immediately surrounding sites. Specifically, I asked two

questions: (i) is pollinator community composition different at sites surrounded

only by crops, compared to sites surrounded by a combination of crops and

unmanaged vegetation? and (ii) is there a relationship between abundance or

richness (species or family) of individual pollinator groups and proximity to either

native mallee vegetation or semi-natural grassland?

4.2 Methods

The study area was located in the Sunraysia agricultural region of northwest

Victoria, Australia (approximately -34.47° to -34.83° S; 142.35° to 142.71° E).

The region is dominated by irrigated horticulture and large remnant areas of

native mallee vegetation. Sampling was conducted over 10 days in late August–

early September 2010 during the brief almond flowering period, which lasts

approximately 2-3 weeks per tree. A total of 72 sites within commercial almond

orchards were sampled in two distinct landscapes. One landscape was a

heterogeneous agricultural mosaic (hereafter ‘Complex’) in which I sampled 27

sites within a landscape diameter of 5.3 km. This landscape was characterised by

53

small blocks of diverse organic and conventional farms (mostly tree fruits and

vineyards) and semi-natural grassland, as well as large patches of natural

vegetation (Figure 1a). The other landscape was a homogeneous landscape

(hereafter ‘Simple’) dominated by broadacre monoculture almond plantations,

large patches of natural vegetation, and occasional smaller patches of semi-natural

grassland or other crops (Figure 1b). I sampled 45 sites in this landscape within a

landscape diameter of 14.6 km. Sites in the Complex landscape were all within

blocks of almond trees and sites in the Simple landscape were located in either

blocks of almond trees or in remnant patches of mallee vegetation within almond

orchard interiors or along orchard boundaries (all patch sites within 15 m of an

almond block). The two landscapes were approximately 36 km apart, and the

minimum distance between any two sites within a landscape was 200 m. Complex

landscape orchards were managed by the same family-owned company and had a

living ground cover of herbaceous vegetation. Simple landscape orchards were

managed by the same corporation and had predominantly bare soil throughout

(see Chapter 3). All orchards were managed without insecticide use.

(a) (b)

Figure 4.1 Representative images showing landscape composition in the (a) Complex (diameter = 5.3 km) and (b) Simple (diameter = 14.6 km) landscapes. Natural vegetation is recognisable as patches of speckled grey, such as the right of image (a) and the bottom left of image (b). In the Complex landscape, crop fields are a diverse array of mostly citrus, avocado, almond, walnut, and table grape vineyards. In the Simple landscape, almond crops cover the bottom half of the circle, as well as the top right corner.

Vegetation types around each site were identified during sampling and the

proportional area of each vegetation type within 100 m of each site was calculated

in ArcMap 10 using spatial data of agricultural land uses in the study region

(SunRISE 21 Inc. 2008). There was no physical overlap in 100 m radii around

54

each site. I focused on a small spatial scale, as mounting evidence suggests that

small-scale attributes related to farm-specific management practices can have

more influence over local insect populations than landscape-scale heterogeneity or

diversity (e.g. Collinge et al. 2006; Kovács-Hostyánszki et al. 2011; Paredes et al.

2013). Pollinators may also travel much shorter distances in agroecosystems when

floral resources are plentiful (Zurbuchen et al. 2010; Lander et al. 2011;

Holzschuh et al. 2013), as they were in my study system. I did not calculate

generalised heterogeneity indices, which often mask the influence of vegetation

structure or resource availability (Cale & Hobbs 1994; Tews et al. 2004); instead,

I documented the proportional availability of specific resources and vegetation

types that were ecologically meaningful to pollinator communities.

4.2.1 Insect sampling and identification

To identify potential wild pollinators in my study system, I focused on individuals

from the taxonomic orders Hymenoptera (non-Apis bees, wasps) and Diptera, as

these taxa are common pollinators in many natural and agricultural systems

(Armstrong 1979; Kevan & Baker 1983; Ssymank et al. 2008). Information on

wild pollinator species in the study region is limited, but wasps and flies are

known to pollinate native flowers in the area (e.g. Ford & Forde 1976; Horskins &

Turner 1999) and individuals of these taxa were seen visiting almond and native

flowers during sampling. I collected potential wild pollinators at each site using

sets of pan traps and design and placement of traps was standardised across all

sites. Each set consisted of one yellow, one white and one blue bowl filled with

water and a drop of detergent to break surface tension. Bowls were made from

white plastic picnic bowls (Woolworths Home Brand) that were either left white

or painted with UV-bright fluorescent yellow or blue paint (White Knight Paints).

Data collected from each coloured bowl per site were pooled for analysis, so the

effect of trap colour is not considered further in this study. Although pan traps

may underestimate the full composition of pollinator insect assemblages, I used

this method to minimise collector bias (Westphal et al. 2008) and I selected trap

colour and placement to increase the chance that the traps would provide

ecologically meaningful results. Yellow and white were the most common flower

colours at sites (in mallee and almond respectively) and blue traps are commonly

55

thought to attract bees, although this colour has not been widely used for native

bees in Australia (see Appendix 1). At each site, the three bowls were placed

together on open ground in a triangular array, adjacent to a flowering tree and

away from fully shaded areas for the duration of the sampling period. Because

many of my target species overwinter or nest in the ground, leaf litter or ground

cover vegetation (Leather et al. 1993), this protocol was designed to ensure

maximum visibility during the late winter sampling time, when many of these

insects would be in the process of emerging or building nests. Each site was

sampled once during the almond flowering period and traps were set out after

sunrise (7.30-9.30 am) and collected before sunset (4.00-6.00 pm) in the sequence

they were set out, so as to capture maximum insect activity during the warmest

part of the day. Insect collection was carried out on days that were fine, calm, and

sunny or partly cloudy with maximum daytime temperatures between 15 and 20°

C. This sampling protocol was not intended to provide a comprehensive survey of

all insect pollinators in this locality; rather, it was designed to provide a

comparative sample of potential wild pollinators across a very large region, in

particular those using almond orchards in different types of mosaic during the

flowering period (Bennett et al. 2006).

Upon collection, insects were stored in 70 % ethanol and counted and

sorted in the laboratory. Native bees were identified to species or subgenus by a

taxonomic expert (K. Walker, Museum Victoria) and I identified wasp and fly

individuals to family level using taxon specific keys (Hamilton et al. 2006;

Stevens et al. 2006). Managed honey bees (Apis mellifera) were present in both

landscapes during sampling, but were not included in the data as the purpose of

this study was to document wild potential pollinator insects available to almond

crops. Wasp and fly individuals were further sorted into small (0-0.49 cm) and

large (>0.5 cm) bodied family groups for analysis, based on the average thorax-

abdomen length of the individuals in each family group. These size categories

were chosen to identify the most likely candidates for almond pollination, based

on the size of an almond blossom (3-5 cm) and position of its stamens; hence,

individuals over 0.5 cm in length were considered more likely to transfer pollen

while visiting blossoms than smaller bodied individuals. The fly family Syrphidae

(hoverflies) was isolated as a separate taxonomic unit from other fly groups, as

this dipteran family is known to be a common pollinator in agroecosystems

56

(Ssymank et al. 2008). All collected native bee and hoverfly individuals were of

the same size class, so each taxon was analysed as a single group.

4.2.2 Data Analysis

Response variables were abundance counts of each insect group (native bee, small

wasp, large wasp, small fly, large fly, hoverfly) and overall species (native bee)

and family (wasp, fly) richness per site (Table 4.1). All data were analysed

untransformed (O’Hara & Kotze 2010). Richness was represented by the bias-

corrected Chao 1 abundance-based estimator. Because of the differences in

orchard management practices and the lack of landscape-scale replication between

the two landscapes (Complex and Simple), I used a non-metric multidimensional

scaling (NMDS) ordination on the entire data set to identify if there was a

difference in pollinator community composition between the two landscapes. Data

were highly aggregated by landscape type (Figure 4.2), and because I was unable

to determine if differences in insect diversity between the two landscapes were

due to specific landscape structural factors rather than management approaches, I

focused on examining relationships within, rather than across, landscapes. Hence,

data were pooled by landscape type (Complex, Simple) for the remaining

analyses.

First, I assessed if overall pollinator community composition differed

between sites surrounded by crops only and sites surrounded by crops and natural

vegetation. I grouped sites based on their floral resource profile, which was

determined by the combination of vegetation types within 100 m of each site:

temporary – only perennial fruit crops with a single brief flowering period each

year; or permanent – crops plus native vegetation or semi-natural grassland. For

each landscape, I used NMDS plots with Bray-Curtis distances to identify

similarity among sites based on their resource profiles. I then used one-way non-

parametric multivariate analysis of variance with Bray-Curtis similarities

(NPMANOVA; Anderson 2001) to test for differences in overall pollinator

community composition between groups of ‘permanent’ and ‘temporary’ sites in

each landscape. I performed this analysis separately for Hymenoptera and Diptera

groups, as previous studies have shown that flies and wasps can respond

57

differently to landscape or floral characteristics (e.g. Jauker et al. 2009; Robson

2013).

Table 4.1 Average abundance and richness of pollinator insect groups per site and average proportion of each vegetation type within 100 m of each site (± SE), in Complex and Simple landscapes. Asterisks signify in which landscape each variable was significantly greater; no asterisk means there was no difference between landscapes.

Complex

(n = 27)

Simple

(n = 45)

Pollinator

abundance

Native bee 2.26 ± 4.2** 0.33 ± 0.9

Large wasp 0.89 ± 1.2* 0.53 ± 1.6

Small wasp 1.93 ± 1.6** 0.58 ± 1.0

Large fly 57.04 ± 79.1** 7.71 ± 7.2

Small fly 29.96 ± 21.3** 4.49 ± 3.0

Hoverfly 0.15 ± 0.4 0.18 ± 0.7

Pollinator

richness

Native bee (species) 1.2 ± 1.4** 0.36 ± 1.0

Wasp (family) 2.89 ± 4.4** 0.86 ± 1.2

Fly (family) 11.43 ± 6.5** 7.36 ± 4.3

Vegetation

Type

Almond 0.66 ± 0.2 0.55 ± 0.4

Mallee 0.09 ± 0.2 0.23 ± 0.3^

Grass 0.08 ± 0.1 0.04 ± 0.1

Grape 0.10 ± 0.2 0.08+

Fruit 0.10 ± 0.1 0

^ p < 0.09; * p = 0.05; ** p < 0.001 + proportion of ‘grape’ for the Simple landscape is based on one site

I then focused on relationships between individual insect groups and the

proportion of native vegetation or semi-natural grassland around each site, as

these were the most undisturbed of the recorded vegetation types and also

represented the most permanent resource availability throughout the year. I

identified the nature of the relationship (positive, negative or neutral) between

richness and abundance of each insect group and each of the two focal vegetation

types using negative binomial generalised linear models (GLMs). Native bee data

from the Simple landscape, and hoverfly abundances from both landscapes, were

not included in these analyses, due to the very low number of individuals caught

in these groups. I did not employ a model selection procedure to rank multiple

58

different models, because the small sample sizes limited the number of parameters

that could be fit to each model and my goal was to identify the nature of the

relationship with the focal vegetation types, rather than to search for potential

relationships among a suite of predictors. Because some sites in the Simple

landscape were located within small mallee vegetation patches inside almond

plantations, I also examined whether the presence of these patches inside the

plantation influenced the overall composition of pollinator assemblages in the

Simple landscape. To do this, I used a bootstrapping procedure to compare the

Shannon diversity indices of Hymenoptera and Diptera families between almond

block sites and within-plantation mallee sites in the Simple landscape. GLMs

were conducted in IBM SPSS 20 (IBM Corporation 2011) and all other analyses

were conducted in PAST 2.17 (Hammer et al. 2001).

Figure 4.2 Overall pollinator community composition in almond orchards located in Complex (black) and Simple (grey) landscapes within the study region. Different symbols denote individual orchards sampled within each landscape. Stress value for the non-metric multidimensional scaling (based on Bray-Curtis similarities) was 0.12, indicating the ordination was a suitable representation of community composition.

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3

59

4.3 Results

4.3.1 Overall landscape and wild pollinator community patterns

The average proportion of almond trees or semi-natural grassland within 100 m of

each site was similar in both Complex and Simple landscapes, but the average

proportion of native mallee vegetation surrounding sites was higher in the Simple

landscape (Table 4.1). Complex landscape sites were surrounded by a small

proportion of other fruit trees and vineyards, but in the Simple landscape there

were no other fruit trees and only one site had grape vines within 100 m (Table

4.1). Across the whole region, I caught a total of 3117 potential wild pollinator

individuals, including 76 native bees, 126 wasps, and 2915 flies (Appendix A2.2).

Average abundance and richness per site of every pollinator group except

hoverflies was highest in the Complex landscape (Table 4.1).

4.3.2 Relationships between wild pollinator communities and site resource

profiles

NMDS ordination of pollinator communities within each landscape showed that

overall pollinator community composition at sites surrounded by ‘temporary’

floral resource profiles (crops only) was very similar to community composition

at sites surrounded by ‘permanent’ resources (crops + natural vegetation) (Figure

4.3). Results of the NPMANOVA tests for differences in the composition of

Hymenoptera and Diptera communities between the two resource profiles

reflected the overall community similarity. In the Complex landscape, there were

no significant differences in community composition between ‘temporary’ and

‘permanent’ sites for either taxonomic group: Diptera, Total Sum of Squares (SS)

= 4.51, Within-group SS = 4.34, F = 1.00, p = 0.36; Hymenoptera, Total SS =

6.58, Within-group SS = 6.35, F = 0.93, p = 0.46 (Figure 4.4a). In the Simple

landscape, there was no significant difference in the composition of Diptera

communities between sites of each resource profile (Total SS = 4.61, Within-

group SS = 4.50, F = 1.08, p = 0.36), but the composition of Hymenoptera

communities showed a difference between groups of sites (Total SS = 15.43,

Within-group SS = 14.55, F = 2.61, p = 0.05) (Figure 4.4b).

60

Figure 4.3 Ordination of pollinator communities at almond sites within (a) Complex and (b) Simple landscapes based on the availability of temporary (perennial fruit crops only) or permanent (crops plus native vegetation and/or semi-natural grassland) resources within 100 m of the site. Stress values for the ordination are Complex = 0.08, Simple = 0.14.

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4

(a)

Perm

Temp

-0.3

-0.25

-0.2

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25

-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4

(b)

Perm

Temp

61

Figure 4.4. (a) In the Complex landscape, community composition of Hymenoptera (top) and Diptera (bottom) groups was similar between sites surrounded by temporary floral resources and sites surrounded by permanent resources.

0

10

20

30

40

50

60

70

80

Prop

ortio

nal a

bund

ance

(a) SmWasp

LgWas

Bee

0

200

400

600

800

1000

1200

1400

1600

1800

Permanent Temporary

Prop

ortio

nal a

bund

ance

Hov

SmFly

LgFly

62

Figure 4.4. (b) In the Simple landscape, only Diptera (bottom) community composition was similar between sites with permanent and temporary resources.

0

5

10

15

20

25

30

35

40

45Pr

opor

tiona

l abu

ndan

ce(b) SmWasp

LgWasp

Bee

0

50

100

150

200

250

300

350

Permanent Temporary

Prop

ortio

nal a

bund

ance

Hov

SmFly

LgFly

63

4.3.3 Relationships between wild pollinator groups and proximate unmanaged

vegetation

Richness and abundance of individual pollinator groups in each landscape showed

contrasting relationships with the proportion of the two focal vegetation types,

native vegetation or semi-natural grassland, around each site (Table 4.2). In the

Complex landscape, none of the relationships were significant and most pollinator

groups were negatively associated with both vegetation types. The exceptions

were large fly abundance, which had a weak positive relationship with both

vegetation types, and small wasp abundance, which had a weak positive

relationship with semi-natural grassland. In the Simple landscape, small wasp

abundance and richness showed a significant negative association with the

proportion of native vegetation near sites and small wasp abundance and fly

richness were negatively associated with the proportion of grassland around sites.

Richness and abundance of all other groups were positively associated with native

vegetation and grassland, but the only significant relationships were between fly

richness and native vegetation and large wasp abundance and grassland.

In the Simple landscape, 88% of hoverflies (7 out of 8 individuals) and all native

bees were collected within mallee patches inside almond orchards, rather than

within blocks of almond trees (Figure 4.5). Diptera family diversity was higher in

the mallee patches inside almond orchards (H = 2.25) than in almond block sites

(H = 1.99; p = 0.02) but there was no difference in Hymenoptera family diversity

between the two site types (mallee H = 1.47, almond H = 1.67, p = 0.37).

64

Table 4.2 Relationships between pollinator groups and the proportion of unmanaged vegetation within 100 m of each site (parameter estimates with standard error in parentheses). Chi-square goodness-of-fit test (χ2) indicates whether the model is a better fit to the data than the constant-only model. Landscape Response NV χ2 GS χ2

Complex A Bee -0.34 (0.53) 0.53 -0.15 (0.47) 0.11

Lge wasp -2.84 (2.23) 2.71 -0.23 (1.10) 0.04

Sml wasp -1.38 (1.13) 1.94 0.03 (0.89) 0.001

Lge fly 0.01 (0.01) 0.13 0.02 (0.01) 1.65

Sml fly -0.13 (0.10) 2.38 -0.02 (0.07) 0.07

R Bee -0.72 (1.29) 0.37 -1.22 (1.59) 0.71

Wasp -0.53 (0.58) 1.15 -1.57 (1.20) 2.96*

Fly -0.15 (0.26) 0.37 -0.46 (0.52) 1.04

Simple A Lge wasp 1.10 (1.50) 0.57 7.50 (3.97)* 3.78**

Sml wasp -1.86 (1.02)* 3.63** -5.97 (4.37) 2.50

Lge fly 0.10 (0.44) 0.05 1.53 (1.15) 1.90

Sml fly 0.63 (0.40) 2.42 0.99 (1.20) 0.68

R Wasp 0.10 (0.71) 0.02 0.21 (2.25) 0.01

Fly 0.82 (0.42)** 3.58* -1.06 (1.94) 0.29

*p < 0.09; **p = 0.05 NV = native vegetation; GS = semi-natural grassland

Figure 4.5 Abundance of each pollinator group in Simple landscape almond blocks (grey) and native mallee vegetation patches inside orchards (black), as a proportion of the total abundance collected across all sites in the Simple landscape (n = 45).

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Bee Large Wasp Small Wasp Hoverfly Small Fly Large Fly

Prop

ortio

nal a

bund

ance

Almond Block Mallee Patch

65

4.4 Discussion

All pollinator groups, except hoverflies, had higher average abundance and

richness in Complex landscape orchards than Simple landscape orchards. This

result was likely influenced by the different orchard floor management practices

in orchards of each landscape type (Chapter 3) as well as the overall structural

differences between the two landscape types. I was not able to separate the

relative influence of these factors on the pattern seen here; however, weedy

ground cover in crop fields and orchards is often considered to be an important

component of landscape complexity (Roschewitz et al. 2005; Pal et al. 2013;

Paredes et al. 2013). Similar differences in pollinator assemblages have been

found in other studies comparing pollinators between simple and complex

agricultural landscapes (e.g. Steffan-Dewenter et al. 2001; Haenke et al. 2009;

Concepción et al. 2012; Persson & Smith 2013) and further sampling across my

study region would enable a more robust comparison between these two

landscape types.

4.4.1 Complex landscape

None of the potential wild pollinator groups in Complex landscape almond

orchards showed significant relationships with the vegetation types surrounding

sites. This could have been a result of the small sample size, or it could indicate

that environmental factors other than these attributes were having greater

influence over pollinator communities. Large flies were the only pollinator group

that were more abundant at sites located near native vegetation or semi-natural

grassland (Figure 4.4a), although this result was largely driven by the extremely

high abundance of the most common fly family, Tachinidae, at a few sites that

were surrounded by more established grassy/weedy ground cover compared to

other sites. This concurs with other studies that have found a strong association

between increased tachinid abundance and perennial or weedy vegetation cover in

farmland (e.g. Abate 1991; Letourneau et al. 2012). Examination of the raw data

showed that abundance and richness of bees and wasps were generally higher at

sites that were only surrounded by temporary resources, i.e. almond, citrus,

avocado trees and table grape vines. This reflected the negative relationships I

66

identified between individual pollinator groups and the proportion of unmanaged

habitats (native vegetation and semi-natural grassland) surrounding sites. Some of

the other fruit crops were flowering during sampling (although not as abundantly

as almond trees), but diverse native floral resources were also available in native

vegetation. These results suggest that hymenopteran pollinators may have been

more attracted to the abundant and diverse floral resources available in the

Complex landscape farrmland, compared to the patchy flowers available in mallee

vegetation, and other studies have also found bee communities to have a stronger

positive association with flowering crop landscapes than with nearby natural or

semi-natural vegetation (e.g. Westphal et al. 2003; Winfree et al. 2008). Most fruit

crops flower in late winter-early spring when wild pollinator species are still

emerging or in their reproductive stage, so agricultural landscapes dominated by

these mass-flowering crop varieties may have a greater opportunity to support the

resource needs of pollinators during this time (Westphal et al. 2009; Jauker et al.

2012). However, functional connectivity of complementary resources within the

matrix around crop fields is also necessary to sustain beneficial insect

communities in the long-term (Jules & Shahani 2003; Williams & Kremen 2007)

and further sampling of sites in the Complex landscape across larger spatial and

temporal scales would provide more information on how wild pollinators use the

heterogeneous landscape around almond orchards after almond trees cease

flowering.

4.4.2 Simple landscape

In the Simple landscape, most pollinator groups showed positive relationships

with the two unmanaged vegetation types, native mallee vegetation or semi-

natural grassland, surrounding sites. Small wasp abundance had a negative

relationship with both these vegetation types, but both unmanaged vegetation

types were negatively correlated with the proportion of almond trees around sites

in the Simple landscape (almond/mallee, r = -0.95, p < 0.001; almond/grassland, r

= -0.55, p < 0.001), so the negative relationship seen here was most likely a

reflection of the high small wasp abundances collected at interior almond sites

more than 100 m from unmanaged vegetation. This result is surprising, due to the

very low plant diversity and structural complexity throughout orchards, which

67

wasps are often strongly associated with (Stephens et al. 2006; Arnan et al. 2011;

Batista Matos et al. 2013). My results could have been influenced by placing pan

traps at ground level, as samples could have been dominated by wasp species that

inhabit leaf litter around tree bases, many of which are small parasitic species

(Austin et al. 2005; Pucci 2008); however, the most common parasitic family

groups I collected (e.g. Diapriidae, Pteromalidae) were more abundant inside

almond orchards than in the within-orchard mallee patches, despite identical

sampling protocols, indicating that these species might prefer the resources

available in almond blocks. These species are unlikely to be useful for almond

blossom pollination, due to their small body size, but they could be important

post-pollination control agents for almond pests (Freeman Long et al. 1998; Eilers

& Klein 2009) and there is scope for further investigation into the distribution of

these wasps in Australian orchards. The relationship between large wasp

abundance and the proportion of semi-natural grassland around sites was the

strongest significant relationship identified in the Simple landscape. Some large

wasp species may use grassland areas for nesting, and other studies have also

found positive relationships between wasp assemblages and areas of grassland

(e.g. Krewenka et al. 2011). However, in my study the majority of sites within

100 m of grassland were also sites located in remnant mallee patches within

plantations and more sampling is necessary to isolate the relative influence of

these two vegetation types.

Interestingly, all bee individuals and all but one hoverfly individual caught

in the Simple landscape were collected within the mallee patches inside orchard

boundaries, rather than in almond blocks themselves. This shows that these key

wild pollinator taxa are present in the landscape, but environmental factors within

the plantations could be limiting their dispersal into the homogeneous interiors. It

is possible that catches of potential pollinators within orchards were low because

the abundant floral resources diminished the effectiveness of the pan traps (Baum

& Wallen 2011); however, the sampling method used in the Simple landscape was

identical to that used in the Complex landscapes, where numerous individuals

were caught within almond blocks, so the low pollinator abundance inside

monoculture plantations was more likely to be influenced by habitat attributes

common to these plantations. Although some studies have found pollinator

abundance to increase in mass-flowering monoculture crop fields, most of these

68

study systems were ground-level or mid-height crops like canola or arable

crops/pasture (e.g. Westphal et al. 2003; Jauker et al. 2009; Holzschuh et al.

2013). In contrast, studies of wild pollinators in intensively-managed tree crops

have found lower pollinator abundance or richness in homogeneous orchard

interiors (e.g. Klein et al. 2012; Marini et al. 2012). Many pollinator species often

prefer to forage in open areas dominated by early successional vegetation

(Kremen et al. 2007; Robinson & Lundholm 2012) and this could explain why

pollinators respond positively to field crop monocultures and negatively to tree

monocultures. Hoverflies and wild bees can also show strong positive

relationships with natural or unmanaged vegetation in intensively-managed

monoculture agriculture landscapes because resource richness is relative to the

local context; therefore, natural vegetation in homogeneous landscapes functions

as resource-rich patches that provide a more obvious contrast with adjacent

broadacre monoculture crop fields than they would with the crop diversity present

in heterogeneous landscapes (Kohler et al. 2008; Haenke et al. 2009). In

particular, Kohler et al. (2008) found that wild bees may not move more than 25-

50 m from favourable habitat patches if the surrounding matrix does not provide

necessary resources. Hence, the establishment of natural vegetation ‘reservoirs’

within farm boundaries to support pollinator communities (Brosi et al. 2008) may

not have much success in increasing the provision of essential unmanaged

pollination services if the farm itself remains an inhospitable environment to

pollinators.

4.4.3 Conclusions and management implications

The results presented here suggest that more intensive management across

agricultural landscapes increases the need for unmanaged habitats providing

resource refuges for potential wild pollinators and other beneficial insects. This

study was limited by the number of suitable replicate orchards available in the

study region, and by the short sampling time, but the findings agree with other

studies that have found patch-scale attributes, like farm management intensity and

proximity of suitable resources, influence within-field pollinator communities

with greater magnitude in homogeneous landscapes compared to complex,

heterogeneous landscapes (e.g. Kleijn & van Langevelde 2006; Kohler et al. 2008;

69

Rundlöf et al. 2008; Persson & Smith 2013). Similar studies of tree fruit crops

have found that generalist wild pollinators are often frequent visitors to orchards

in the United States and Europe, but their presence is mostly limited to parts of the

orchard with proximity to natural or semi-natural habitat patches that can sustain

insect resource needs before and after orchard trees are in flower (Watson et al.

2011; Klein et al. 2012; Marini et al. 2012). Thus, agricultural landscapes

dominated by tree crop plantations have the potential to support diverse wild

pollinator assemblages, but less intensive management practices adopted at the

landscape or regional scale are necessary to provide permanent spatial and

temporary resources that are essential for wild pollinator communities to persist

(Williams & Kremen 2007; Holzschuh et al. 2008). These findings have

implications for conservation and agricultural production. Irrigated agricultural

areas in arid landscapes, like the almond plantations in my study system, have the

ability to support diverse wild insect communities, potentially providing essential

resources during extended dry periods. However, intensive management of these

areas reduces their conservation potential, as many wild pollinators and other

beneficial insects may be unable to persist inside monoculture fields and orchards.

This could affect the long-term composition of local insect assemblages and,

therefore, the availability of insects that provide essential pollination services to

pollinator-dependent crops and native flowering plants. Functional diversity has

more influence over community assembly and ecosystem functions than

abundance or richness of a particular species group (Mayfield et al. 2010; Cadotte

et al. 2011; Lavorel et al. 2011); hence, studies like mine that focus on functional

groups of beneficial insects in broadacre monoculture landscapes, rather than only

commonly-recognised taxa like wild bees, are necessary to identify how structural

homogeneity in agroecosystems can influence the provision of ecosystem

services.

70

71

Chapter 5: The influence of a semi-arid woodland—almond plantation edge

on wild pollinator communities

Abstract

Agricultural landscape elements, such as field edges, are not always a barrier to

insect movement but can influence their distribution and dispersal behaviour. I

investigated spatial and temporal patterns in wild pollinator (fly, wasp and native

bee) distribution across an ecotone between natural mallee woodland and

monoculture almond plantations in southern Australia, during the critical almond

flowering period. This is the first study of variation in pollinator community

distribution on both sides of the habitat boundary between natural woodlands and

flowering tree crop plantations. I also use SHE analysis (species richness,

diversity and evenness) to identify how the edge influenced pollinator community

structure across the flowering season. I show that the spatial distribution and

structure of pollinator communities can vary across a habitat boundary with

temporal changes in resources. The physical habitat boundary was not a barrier to

wild pollinators, but appeared to be influencing their movement into plantation

interiors. Results suggest that pollinator communities were more likely influenced

by vegetation heterogeneity and phenological changes in resources, than by the

physical presence of the edge. My findings contribute to knowledge of edge

responses by insects in managed landscapes and may also encourage the adoption

of ecological management practices in commercial plantations of pollinator-

dependent crops.

*This chapter has been published as Saunders ME, Luck GW (2014) Spatial and

temporal variation in pollinator community structure relative to a woodland—

almond plantation boundary. Agricultural and Forest Entomology, doi:

10.1111/afe.12067.

72

5.1 Introduction

Wild insect pollinators are keystone organisms and, globally, are being

increasingly affected by the intensification of agriculture and the modification of

natural habitats and landscapes (Winfree et al. 2011; Wratten et al. 2012).

Landscape structure strongly moderates biodiversity patterns and ecological

processes (Tscharntke et al. 2012) and boundaries between farmland and natural

vegetation can be dynamic ecosystems themselves, influencing insect populations

and trophic interactions across time and space (Hunter 2002). The terms

‘boundary’, ‘ecotone’, and ‘edge’ may be defined in different ways (van der

Maarel 1990; Cadenasso et al. 2003; Yarrow & Marín 2007); in this paper, I use

the terms ‘boundary’ or ‘edge’ to refer to a physical point of change between two

adjacent vegetation types, and the term ‘ecotone’ to refer to a diffuse zone

surrounding the boundary on both sides, where the potential for ecological

influence from edge processes exists. Anthropogenically-created edges can have

different dynamics to natural edges – they have a sharp gradient, little or no

transition zone between the plant communities sharing the edge (Farina 1998),

and are regularly influenced by human activities (Dutoit et al. 2007; José-María et

al. 2010). These types of edges are characteristic of agroecosystems, where each

field, paddock or plantation is often delineated by fence lines or roadways, and

they are becoming more common worldwide as intensively cultivated landscapes

expand (Blitzer et al. 2012). Therefore, understanding how wild pollinators are

influenced by physical boundaries in agroecosystems contributes valuable

information to pollinator conservation programs and the ecological management

of agroecosystems.

Although anthropogenic edges in agroecosystems can affect the movement

of birds and mammals (e.g. Yahner 1988; Luck et al. 1999; King et al. 2011), they

may not always be a physical barrier to insect movement (Forman & Godron

1986). However, they can still affect insect life cycles, ecological interactions,

and, ultimately, the distribution and dispersal of insect communities across the

landscape (Fagan et al. 1999; Haslett 2001; Hunter 2002). Many studies that have

investigated wild pollinators near natural/agricultural boundaries have focused on

pollinator communities within particular crops, and therefore have only sampled

on the farmland side of the boundary (e.g. Chacoff & Aizen 2006; Jauker et al.

73

2009; Carvalheiro et al. 2010). However, the ecological effects of edges are not

one-sided and, because ecological processes in the surrounding matrix can

influence edge responses in conjunction with processes occurring in the focal

patch, ‘two-sided’ sampling provides a more rigorous estimate of how an edge

might influence adjacent ecosystems (Ries et al. 2004; Ewers & Didham 2006;

Fonseca & Joner 2007). Many insects found in farmland will travel there, often

temporarily, from nearby native vegetation (Blitzer et al. 2012) and, as in some

natural ecosystems, may rarely be confined to one side of a landscape boundary

(Dangerfield et al. 2003). More studies that specifically quantify insect abundance

or diversity on both sides of farmland/natural vegetation edges will improve

understanding of how these edge types influence insect communities (sensu

Forister 2009; Campbell et al. 2011; Lucey & Hill 2012).

Empirical studies of natural edges suggest that a species' edge response is

often related to the specific ecological and landscape context (Campbell et al.

2011; Macfadyen & Muller 2013). Insect movements at boundaries may be

influenced by patch size or shape (With & Pavuk 2011), boundary permeability or

complexity (Haslett 2001; Stasek et al. 2008), or an interaction between these

variables (Collinge & Palmer 2002), and different insect taxa can show

contrasting responses to the same edge (Olson & Andow 2008; Pryke & Samways

2012). For example, Jauker et al. (2009) found that wild bee abundance declined

while hoverfly abundance increased along field margins as the distance to natural

habitat increased, because of the differences in dispersal behaviour and resource

requirements between the two taxa. Individual species responses ultimately affect

community composition at the edge, although very few studies have considered

community-level responses to habitat edges (Ries et al. 2004; Pryke & Samways

2012).

Spatial patterns are often linked to temporal variation within the system,

although these relationships are also contextual (Preston 1960; MacArthur &

Wilson 1963; Adler & Lauenroth 2003; Soininen 2010). The interaction between

spatial and temporal variation of species distributions is especially pertinent when

considering ecosystem service provision in agroecosystems (Kremen 2005;

Bianchi et al. 2010; Holland et al. 2011), where temporal changes in relevant

farmland attributes, such as floral resources, often happen abruptly and with

greater magnitude than in natural ecosystems. For example, floral resources in a

74

large monoculture crop field will be spatially homogeneous, available only for a

very short time period, and saturate a large area of land, compared to a natural

ecosystem that may support diverse and patchy floral resources throughout the

year with constant spatial variation of flower location. Transient, large-scale

mass-flowering events common to farmland have the potential to lure wild

pollinators away from natural habitats (the ‘Circe Principle’, Bartomeus &

Winfree 2011; Lander et al. 2011) and this can have negative impacts on native

plant pollination as well as insect reproduction, especially for those species that

rely on permanent, abundant floral resources (Westphal et al. 2009; Holzschuh et

al. 2011). Thus, examining the distribution patterns of ecosystem service

providers (e.g. pollinators) in agroecosystems will benefit from considering both

spatial and temporal variation of species patterns, especially in relation to the

relevant life-stage of the crop (e.g. flowering period; sensu Mesa et al. 2013;

Peters et al. 2013).

Here, I investigate spatial and temporal patterns of wild pollinator

abundance and diversity across an ecotone between two vegetation types in

southern Australia: natural mallee woodland and monoculture almond plantations

(Prunus dulcis, Mill.). Almond is a 100 % pollinator-dependent crop, yet there is

currently no available information on potential wild pollinators available to

almond crops in Australia (but see Saunders et al. 2013). I did not quantify

pollination efficiency in this study, but I use the term 'pollinator' to refer to insects

in the orders Hymenoptera and Diptera to maintain consistency with the

pollination literature. Wasps and flies are important but understudied pollinators

in a number of natural and agricultural systems (Kevan & Baker 1983; Larson et

al. 2001; Ssymank et al. 2008), particularly in Australia (Armstrong 1979;

Pickering & Stock 2003) and individuals of these orders were seen visiting

flowers during sampling. Because of the limited information on pollinator taxa in

my study system, I focus on documenting spatial and temporal patterns to initiate

future research and hypothesis generation (e.g. exploring the drivers behind these

patterns), as these patterns are fundamental to understanding ecological processes

and ecosystem function (Turner 1989; Levin 1992). I also use SHE analysis

(Buzas & Hayek 1996) to examine how the habitat edge influenced pollinator

communities by identifying changes in community structure along the physical

75

woodland—plantation gradient and comparing these taxonomic boundaries with

the location of the habitat boundary.

5.2 Methods

5.2.1 Study area and sampling

The study area is located in the Victorian Murray Mallee Bioregion, in the

neighbouring districts of Wemen and Liparoo, northwest Victoria, Australia (34°

S 142° E). The Murray Mallee Bioregion is semi-arid and average annual rainfall

in the study area is approximately 315 mm (Bureau of Meteorology 2013).

Vegetation is dominated by Eucalyptus, Acacia, Callitris and Casuarina short

trees (approx. 3-12 m) and tall shrubs, while the understorey is highly variable,

with patchy ground cover vegetation typically including hummocks of Triodia,

halophytic plants, and a range of herbs and annual species (Wood 1929; Cheal &

Parkes 1989).

I sampled within two conventional monoculture almond plantations

(Liparoo, Wemen) and two mallee woodlands. To minimise the effects of

different edge orientations or contrasts (Ries et al. 2004) I chose plantations in the

same geographical region with large patches of adjacent native mallee vegetation

that had similar orientation and structure (Figure 5.1). Both mallee—almond

boundaries were aligned east-west, with mallee vegetation on the southern side of

the edge and almond plantation on the northern side, and both were delineated by

a fence line and a two lane road approximately 5 m wide (an unpaved farm road at

Liparoo and a paved municipal road at Wemen). Liparoo is adjacent to the

Annuello Flora and Fauna Reserve (hereafter Annuello) and Wemen is adjacent to

an unnamed patch of remnant mallee vegetation (hereafter Collins). Both

plantations are broadacre monocultures, intensively managed by the same

company, using the same practices. The linear distance between the two

plantations is approximately 7.5 km. Plantations differed in their age and size, and

mallee patches differed in their size and management history (Appendix 4 Table

A4.1).

76

Figure 5.1 The study area in northwest Victoria, Australia, showing the sampled mallee/almond edges at Liparoo (left) and Wemen (right) plantations – mallee vegetation patches are on the south side of each edge and almond plantations are on the north side of the edge.

Insects were collected on three occasions in 2011, coinciding with the flowering

of the almond crop in late winter – before (July), during (August) and after

(September) almond flowering, which lasts approximately 3-4 weeks. July is

typically the coldest month of the year in my study region, and deciduous almond

trees were completely bare at this time (Figure 5.2). Almond flowering occurred

across the entire plantation in August, with each tree still predominantly leafless,

but completely covered in blossom (approx. 50,000 individual flowers per tree).

In September, the first month of spring, the majority of almond blossoms had

dropped and trees were fully leaved. In mallee woodlands, phenological changes

were much less obvious, with evergreen plants and trees and diverse, patchy floral

resources available in all months (Figure 5.2). Hence, I present a novel assessment

of temporal changes in pollinator communities across a mallee—almond ecotone,

in the context of a very brief, but substantial, resource pulse with the potential to

77

influence pollinator distribution, i.e. landscape-scale mass-flowering of

monoculture almond plantations.

Figure 5.2 Representative photographs of (top left) almond plantations in July; (top right) plantations in August; (bottom left) plantations in September; (bottom right) mallee woodlands.

Ries et al. (2004) suggested that the depth of an edge's influence on invertebrates

can be up to 100 m; hence, we considered 300 m from the edge to be an

acceptable distance for interior sites, given average pollinator foraging ranges,

which can be highly variable among taxa and under different environmental

conditions (Nielsen et al. 2012; Essenberg 2013; Jha & Kremen 2013). Non-Apis

bees also exhibit small-scale responses to floral resources, especially when

resources are at higher densities (Greenleaf et al. 2007; Jha & Vandermeer 2009;

Essenberg 2013). To account for this, I sampled more sites in plantations than in

mallee woodlands, because we predicted that changes in pollinator communities

might occur more rapidly on the almond side of the edge. On each sampling

occasion, I sampled 50 almond sites (25 in each plantation) and 18 mallee sites

(nine in each mallee patch) along parallel transects perpendicular to the edge. The

number and location of transects were chosen a priori from satellite maps and

were limited by access and cadastral boundaries. Sites were located at 10 m

(edge), 40 m (intermediate) and 300 m (interior) on both sides of the edge. On the

plantation side of the edge only, I sampled additional intermediate distance classes

78

at 70 m and 100 m. Thus, each plantation transect consisted of five sites and each

woodland transect consisted of three sites. All transects were at least 200 m apart

and the transect was considered the replicate in time and space.

Pan traps may underestimate some taxa, but I used this method to

minimise collector bias that can arise from other direct sampling methods

(Westphal et al. 2008; Popic et al. 2013) and designed bowl colour and placement

to increase the likelihood that data would be ecologically meaningful (see

Appendix 1). I used UV-bright yellow and white bowls, as white was the

predominant flower colour in plantations and yellow the most common in

woodlands. Because of the winter sampling time, traps at each site were open for

8-10 hours in each month to capture maximum insect activity during the warmest

part of the day. Traps were placed on the ground next to a flowering tree for

maximum visibility and were set just after sunrise and collected just before

nightfall in the order they were placed. Insects were stored in 70% ethanol and

transported to the laboratory, where they were counted and sorted into taxonomic

order (Diptera (flies) and Hymenoptera (native bees and wasps)). I identified

native bees to species (where possible, although most Lasioglossum individuals

could only be classified to subgenus) and flies and wasps to family level using

specialised identification keys (Hamilton et al. 2006; Stevens et al. 2007) along

with a reference collection provided by a taxonomic expert (K. Walker, Museum

Victoria).

5.2.2 Data analysis

I analysed differences in monthly pollinator abundance and richness between

vegetation types, and compared how insect community structure changed relative

to the edge. The insect abundance data were overdispersed ecological counts and

were analysed untransformed, as recommended for these type of data (McArdle &

Anderson 2004; O’Hara & Kotze 2010; Warton & Hui 2011). All statistical

analyses were conducted in PAST 2.17 (Hammer et al. 2001) using raw

abundance counts, a species richness index (hereafter simply 'richness'), or other

diversity measures as described below. I calculated richness with the Menhinick

index, which takes the number of individuals per species into account and is

robust to unequal sample sizes (Menhinick 1964). Analyses on the native bee data

79

were conducted on September counts only as I caught zero individuals in July and

4 individuals in August.

I first looked at overall distribution patterns of potential wild pollinators

across the whole study area, regardless of month, and then examined spatial and

temporal trends in more detail, specifically how monthly populations changed in

each vegetation type and how spatial distribution changed across time. To

examine overall distribution patterns relative to the sampled distance gradient, I

compared community similarity between distance classes, as this approach is

recommended for comparing site diversity relative to a distance gradient

(Tuomisto & Ruokolainen 2006; Soininen 2010). I calculated average Bray-Curtis

similarity coefficients for overall community similarity (total Diptera and total

Hymenoptera) at each possible pair of distance classes, using all combinations of

individual sites. I used linear regression (ordinary least squares) to test whether

community similarity declined as geographic distance between distance classes

increased along the transect (the 'distance-decay' phenomenon; Nekola & White

1999; Soininen 2010). The fit of the model was verified by examining residuals

for autocorrelation and heteroscedasticity. I then tested whether heterogeneity of

vegetation type contributed to heterogeneity of pollinator communities by

grouping similarity coefficients based on the three possible combinations of

vegetation pairs (mallee/mallee, almond/almond, and mallee/almond) and

comparing the mean similarity between each group using a standard t-test. I then

examined monthly turnover in abundance and richness of each pollinator group

per site, relative to the distance gradient. Wilcoxon signed-rank tests were used to

look for significant differences in fly and wasp richness and abundance between

sequential months (July-August, August-September) at each distance class, within

each vegetation type (native bees were not included in this analysis because nearly

all individuals were caught in September).

To identify whether the physical edge was influencing pollinator

community composition, I separated data from each woodland/plantation pair

(Annuello—Liparoo and Collins—Wemen) to account for differences in patch

area (see Appendix 4 Table A4.1). I tested whether changes in Hymenoptera and

Diptera community composition along each ecotone gradient coincided with the

physical boundary between vegetation types using SHE analysis (Buzas & Hayek

1996; Hayek & Buzas 1997). Interior (300 m) mallee sites were designated as the

80

starting point of the ecotone, under the assumption that mallee woodland was the

'source' habitat for native insects in the study area, and I quantified the gradient as

follows: -300 m, -40 m, -10 m, 10 m, 40 m, 70 m, 100 m, 300 m, with negative

values indicating mallee woodland sites and positive values indicating almond

plantation sites.

I analysed bees and wasps separately because exploratory analysis showed

that the presence of an extreme outlier at one distance class in the bee data skewed

the analysis and masked the patterns in the wasp data set. SHE analysis examines

the relationship between three core measures of biodiversity, species richness (S),

diversity (H) and species evenness (E) as they accumulate through quadrats,

transects or time. Mathematically, this technique is similar to other methods for

analysing species richness, such as species accumulation curves (Buzas & Hayek

1996; Buzas & Hayek 1998); hence, it can also be applied to assemblages

identified to higher taxonomic levels, e.g. genus or family (Gotelli & Colwell

2011). SHE analysis has two main applications: as an investigation of the

diversity and underlying distribution of the sample population (Buzas & Hayek

1996; Magurran 2004); and as a robust measure for analysing zonation in

community structure along a spatial or temporal gradient (Buzas & Hayek 1998;

Osterman et al. 2002; Magurran 2004). SHE analysis has been used mainly in the

analysis of marine foraminiferal assemblages and palaeo-communities (e.g. Buzas

& Hayek 1998; Osterman et al. 2002), but has also been applied to a range of

terrestrial systems (e.g. Small & McCarthy 2002; Leponce et al. 2004; Parra-H &

Nates-Parra 2012). Because the analysis follows the sampled spatial or temporal

gradient in a sequential fashion, rather than analysing pairwise comparisons

between all pairs of samples, it is recommended for identifying boundaries within

faunal communities over ordination or clustering techniques (Osterman et al.

2002), which can mask the data's true complexity, overlook important

environmental or temporal relationships and confound dispersion and location

effects in the data (Osterman et al. 2002; Warton 2008; Warton et al. 2012).

SHE analysis assumes that a single 'closed-system' community has a log-

series distribution (Buzas & Hayek 2005), for which a plot of lnE vs lnN should

exhibit a decreasing linear trend (Buzas & Hayek 1998). Where the line departs

from this trend, a statistically significant change in the faunal assemblage has

occurred and the communities on either side of the break point can be assumed to

81

be separated in time or space, whichever is the focus of the sampled gradient

(Buzas & Hayek 1998; Osterman et al. 2002; Wilson 2008). I calculated S, H and

E for accumulated abundances (N) in each sequential sample along each ecotone

gradient (600 m), and plotted lnE vs lnN for the entire transect (following Buzas

& Hayek 1998). I examined the initial lnE vs lnN plot, and when a break in the

linear trend occurred, the samples up to and including the break point were

discarded and lnE vs lnN for the remaining samples was replotted. This process

continued until all samples were accounted for. The resulting wasp, fly, and bee

'abundance zones' were then compared with the distance classes along the

physical mallee—almond ecotone gradient.

5.3 Results

5.3.1 General patterns

I caught a total of 2394 fly individuals in 20 family groups, 336 wasp individuals

in 23 families, and 171 non-Apis bee individuals in 2 genera (Lasioglossum

(Chilalictus) and Homalictus (Hymenoptera: Halictidae)) across the whole

sampling period (Appendix 2 Table A2.3). Per trap site, average bee abundance

was highest at Annuello (mean 8 ± 13.3), average wasp abundance was highest at

Collins (mean 8.4 ± 10.04) and average fly abundance was highest at Wemen

(mean 21.42 ± 43.02). Average fly richness was highest at Annuello (mean 1.92 ±

0.73), and average wasp (mean 1 ± 0.54) and bee (mean 0.65 ± 0.5) richness were

highest at Collins. Five Hymenoptera families (Halictidae, Ichneumonidae,

Mymaridae, Pteromalidae, Scelionidae) and seven Diptera families

(Calliphoridae, Chironomidae, Chloropidae, Dolichopodidae, Heleomyzidae,

Phoridae, Tachinidae) were collected at every distance class along the ecotone

(Appendix 4 Table A4.2).

5.3.2 Distance decay and monthly turnover

Pollinator community similarity between sites declined significantly as the

distance between them increased, but the slope of the regression was very weak

(Figure 5.3). Heterogeneity of vegetation types influenced community similarity

82

between sites. On average, almond/almond site pairs had higher community

similarity (mean 0.75 ± 0.14) than either almond/mallee (mean 0.69 ± 0.18; F =

1.59, t = 7.25, p < 0.001) or mallee/mallee site pairs (mean 0.69 ± 0.17; F = 1.53, t

= -3.45, p < 0.001). Similarity between mallee/mallee site pairs was not different

to almond/mallee site pairs (F = 1.04, t = -0.25, p = 0.8) (Figure 5.4).

Figure 5.3 Distance-decay graph showing the decline in pollinator community similarity with linear distance between sites along the 600 m mallee—almond ecotone.

Average bee and wasp abundance and richness increased gradually across the

sampling period, except for wasp richness in mallee woodland, which declined

slightly in August before another increase in September. Average fly abundance

increased each month in almond plantations, but fly abundance in mallee and fly

richness in both vegetation types peaked in August and declined again September.

Fly and wasp richness and abundance along the mallee section of the ecotone did

not change significantly between months (Figure 5.5). On the plantation side of

the boundary, there were no significant increases in wasp richness and abundance

from July-August at any distance class, but both richness and abundance increased

r² = 0.3945, similarity = -0.0002(distance) + 0.74370.5

0.55

0.6

0.65

0.7

0.75

0.8

0 100 200 300 400 500 600 700

Bra

y-C

urtis

sim

ilarit

y

Distance (m)

83

significantly from August to September at 10 m, 70 m and 300 m (Figure 5.5a).

Fly abundance at all distance classes in almond and richness at 10 m, 100 m and

300 m increased significantly from July to August and then remained similar from

August to September (Figure 5.5b).

5.3.3 Influence of the edge

The SHE analysis identified changes in community structure (community zones)

relative to the physical distance gradient. At Collins—Wemen, there were three

distinct community zones, while only two were apparent at Annuello-Liparoo

(Table 5.1). Wasp and fly community composition at Collins—Wemen changed at

the boundary in August and September, and then again around 100 m inside the

plantation, while the native bee community changed on the mallee side of the

boundary in September (between 40 and 10 m). In July, wasp community

composition in the Collins (mallee) interior remained similar across the boundary

and into Wemen (almond), changing between 70 and 100 m inside the plantation,

and fly community boundaries were identified between 10 and 40 m and after 100

m on the almond side. At Annuello—Liparoo, wasp and fly communities also

changed near the physical boundary in September, while native bee communities

changed structure further inside the plantation (between 40 and 70 m). In July and

August, wasp and fly community boundaries were identified on the almond side

of the edge. Wasp zones changed between 70 and 100 m inside the plantation in

both months, and fly zones changed around 40 m inside the plantation (Table 5.1).

84

Figure 5.4 Mean (± SD) pollinator community similarity (Bray-Curtis) between site pairs of each combination of vegetation type. Lower-case letters above bars indicate whether each group of site pairs was significantly different from other groups (p < 0.01).

a b a

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Mallee-Mallee Almond-Almond Almond-Mallee

Aver

age

polli

nato

r com

mun

ity s

imila

rity

85

Figure 5.5 (a) Monthly average wasp abundance and richness at each distance along the ecotone throughout the sampling period (July-September). Asterisks indicate a significant increase/decrease from the previous month (p < 0.05).

* *

*

0

0.5

1

1.5

2

2.5

3

3.5

4

Aver

age

abun

danc

e

(a) Jul

Aug

Sep

* *

*

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

-300 -40 -10 10 40 70 100 300

Aver

age

richn

ess

Distance along ecotone (m)

86

Figure 5.5 (b) Monthly average fly abundance and richness at each distance along the ecotone throughout the sampling period (July-September). Asterisks indicate a significant increase/decrease from the previous month (p < 0.05).

*

* * *

*

0

5

10

15

20

25

30

Aver

age

abun

danc

e(b) Jul

Aug

Sep

**

*

0

0.5

1

1.5

2

2.5

3

-300 -40 -10 10 40 70 100 300

Aver

age

richn

ess

Distance along ecotone (m)

87

Table 5.1 Community zones identified from SHE analysis for wasp, fly and native bee communities, relative to the distance along the physical ecotone at each location (negative values indicate mallee sites and positive values indicate almond sites). Shading indicates the different community zones. Only September community zones are shown for native bees. Wasp Distance along ecotone gradient (m)

Ann—Lip -300 -40 -10 10 40 70 100 300

July

August

September

Coll—Wem

July

August

September

Fly Distance along ecotone gradient (m)

Ann—Lip -300 -40 -10 10 40 70 100 300

July

August

September

Coll—Wem

July

August

September

Native beeDistance along ecotone gradient (m)

-300 -40 -10 10 40 70 100 300

Ann—Lip

Coll—Wem

Ann—Lip = ecotone between Annuello (mallee) and Liparoo (almond); Coll—Wem = ecotone between Collins (mallee) and Wemen (almond)

88

5.4 Discussion

5.4.1 Distance decay

Across the whole sampling period Hymenoptera were, on average, more abundant

and more diverse in mallee woodland and Diptera were more abundant in almond

plantations but more diverse in mallee woodland. There was an overall distance

decay of pollinator community similarity between distance classes along the

sampled distance gradient. Nekola & White (1999) suggest two possible causes

for this phenomenon: a decrease in environmental/habitat similarity with distance,

which influences niche availability; or changes in spatial configuration and

landscape context across space and time, which affect species' dispersal. These

factors are inherently linked and it can be difficult to separate the underlying

causes of observed variation in a distance decay model (Nekola & White 1999).

Mallee woodland is highly-complex and heterogeneous in structure compared to

the monoculture plantations, where herbicides are used to maintain bare ground

throughout and blocks of almond trees are nearly identical in row width and tree

height. I found that the homogeneity in plantations was reflected in pollinator

communities – similarity of insect communities between site pairs was highest

among almond sites, while variation between pollinator communities within

mallee vegetation was just as great as it was between the two different vegetation

types. Habitat heterogeneity and complexity increase niche availability, and can

also have a significant influence over invertebrate communities, trophic webs and

overall ecosystem function (Srivastava 2006; Randlkofer 2010a), particularly in

agroecosystems (Altieri 1999; Letourneau et al. 2011). Lower rates of distance

decay, as seen in the regression analysis here, can occur because of minimal

changes in the environment along the sampled gradient or because the focal taxa

are highly mobile and less affected by landscape barriers such as edges (Nekola &

White 1999). This latter explanation is most likely for my results, as vegetation

structure and heterogeneity changed significantly along the ecotone and

influenced community similarity of my focal taxa, highly-mobile insect

pollinators. Further sampling at more distance classes along the ecotone,

combined with a comparison of distance decay rates between insect groups,

89

vegetation types and months would yield more information on the significance of

the distance-similarity relationship in this study system.

5.4.2 Temporal variation

Each pollinator group showed different rates of temporal turnover, reflecting the

variation in resource needs of individual taxa. Wasps showed the weakest

responses in my system, despite evidence that some wasps may be more sensitive

to anthropogenic change than bees or other insects (Batista Matos et al. 2013).

Wasp abundance and family richness increased steadily, and at a lower rate, in

both vegetation types from July through to September, but were also consistently

higher in woodlands than plantations. Fly family richness and abundance showed

much stronger fluctuations throughout the sampling period, and flies were also

more abundant overall in my study system. Dipteran species can be more cold-

tolerant than hymenopterans and are often the most frequent flower-visitors in

colder climate zones and higher latitudes (Elberling & Olesen 1999; Pickering &

Stock 2003; Brown & McNeil 2009) so the low abundances of hymenopterans

most likely reflects the season I sampled in. Native bees were absent in samples in

July and rarely collected in August during almond flowering, even though

sampling from 2010 showed that native bees are typically present in the study

region during these months, including in plant-diverse almond orchards managed

with lower intensity (Chapter 3). Many non-Apis bees need access to multiple

heterogeneous habitats throughout the year (Winfree et al. 2011; Mandelik et al.

2012) so permanent spatial homogeneity and lack of floral diversity within

perennial monocultures could prevent bees from establishing inside plantations,

regardless of the resources provided by the short-term floral pulse, but more

sampling is needed to verify this effect.

Although pollinator abundance increased naturally across the sampling

period in response to seasonal warming, I also found evidence that some insects

may have been influenced by the phenological changes occurring in almond

plantations. In July, mallee sites had higher fly and wasp abundance and richness

than almond plantations, most likely because of greater availability and

heterogeneity of resources in woodlands. Pollinator abundance was also higher in

the mallee interior than the almond interior in July, despite the mallee having

90

higher structural density at the edge, an attribute that is considered to prevent

movement of some insects into forest interiors (Cadenasso & Pickett 2000). These

results suggest that consistent resource availability beyond an edge may override

the qualities of the edge itself, especially for pollinators (Kohler et al. 2008;

Stasek et al. 2008) or in agroecosystems with high temporal variability of

resources (Rand et al. 2006). During the almond flowering event in August, fly

abundance and richness increased significantly in plantations, and then remained

at similar levels into September. In contrast, wasp abundance and richness did not

increase significantly in the plantations until September – although there was a

small increase in wasp populations in August. This could indicate a delayed

response to the flowering event (Thomson 1981; Ries et al. 2004), which often

occurs with short, substantial resource pulses (Ostfeld & Keesing 2000).

Alternatively, wasp assemblages could have been influenced by the additional

resources available in plantations after flowering. As well as the spent flowers,

most trees were fully-leaved and nut formation had begun, and this coincided with

an increase in herbivorous insects such as planthoppers, aphids (Hemiptera) and

psocids (Psocoptera), which are potential prey or host species for many wasps and

flies (see Chapter 6). Temporal increases in pollinator abundance and richness on

the mallee side of the boundary were not significant, so the increases I observed at

almond sites were more likely to be influenced by the simultaneous changes in

resource availability, rather than seasonal changes in weather conditions.

Although managed European honeybees were present in plantations during the

August sampling event, it is unlikely that the increase in honeybee abundance

significantly influenced my results, due to the abundant floral resources (e.g.

Markwell et al. 1993). As discussed in Chapter 2, evidence suggests that

honeybee and non-Apis pollinator insects can show spatial complementarity in

foraging and nesting locations. In particular, a study in North American almond

plantations found that wild pollinators and European honeybees partitioned

themselves across different parts of the tree while foraging, rather than being

deterred althogether (Brittain et al. 2013).

91

5.4.3 Influence of the edge

The results of the SHE analysis suggested that the physical presence of the edge

was not a barrier for pollinators, but the spatial and temporal changes in

vegetation structure and resources at the woodland/plantation boundary did have

an influence on community structure. Pollinators are highly-mobile, but can also

be sensitive to changes in habitat that do not provide the food or structural

resources they need (Kremen et al. 2007; Kohler et al. 2008). Therefore, edge

responses for these species can be highly-variable, related to contextual attributes

of the studied habitat (Jauker et al. 2009; Azevedo et al. 2011; Coudrain et al.

2013; Macfadyen & Muller 2013) or specific to each taxa, with flies, bees and

wasps all showing differing responses within a single ecological context (Klein et

al. 2004; Jauker et al. 2009; Rader et al. 2011; Batista Matos et al. 2013). I

identified distinct community zones (i.e. significant changes in community

structure or composition) along the mallee/almond gradients, and the boundaries

between these zones shifted across the sampling season and also differed between

plantations. Most importantly, these boundaries did not always reflect the location

of the woodland/plantation edge. Two distinct community zones were identified

along the mallee—almond ecotone between Annuello woodland and Liparoo

plantation. All changes in community structure of each pollinator group occurred

between the physical edge and 100 m inside the plantation (Table 5.1), suggesting

that this could be the maximum distance from the source habitat (mallee) that

many pollinators would travel. For example, the wasp families Ichneumonidae

and Tiphiidae, both common flower-visitors (Kevan 1973) and important

pollinators for some Australian native flowers (Armstrong 1979; Peakall &

Handel 1993), were only found up to 70 m inside the plantation (Appendix 4

Table A4.2). In addition, although both bee species collected at Annuello and

Liparoo were found across the whole ecotone, 83 % of the total individuals were

collected at mallee and almond edge (10 m) sites and almond intermediate (40 m)

sites. This suggests that pollinators living in Annuello were utilising resources in

plantations throughout the flowering season, but environmental factors were

inhibiting their movement beyond 100 m. This hypothesis agrees with Ries et al.’s

(2004) finding that invertebrate responses to edges generally extend up to 100 m

from the actual edge, as well as established knowledge that pollinators,

92

particularly bees, require access to multiple different resources to complete their

life cycle (Kremen et al. 2007; Mandelik et al. 2012).

In contrast to the Annuello—Liparoo ecotone, three distinct community

zones were identified along the Collins—Wemen gradient. All groups, except

July wasp communities, changed structure on the mallee side of the boundary,

between 40 m and the edge, and then again between 70 and 300 m inside the

plantation. This could indicate that some pollinators were congregating near to the

edge, creating a community structure distinct from the interior sites, but this was

not the case with all taxa (Appendix 4 Table A4.2). For example, the bee species

L. (Chilalictus) erythrurum was found at 100 m inside Wemen plantation and this

species was not collected at all along the Annuello—Liparoo ecotone. In addition,

55 % of bee individuals collected at Collins—Wemen were found between 70 and

300 m inside the almond plantation. Because the two mallee/almond edges were

similar in their orientation, geographic location, edge contrast and general

vegetation structure, these results suggest that other ecological factors within the

plantation are influencing pollinator movement into plantation interiors. Both

Wemen and Liparoo plantations have a high proportion of native vegetation in the

surrounding landscape (mallee woodland along the studied edge and floodplain

red gum (Eucalyptus camaldulensis) forests bordering the other side of the

plantation); however, Wemen is much smaller in area than Liparoo (Appendix 4

Table A4.1) and the greatest distance from an edge to plantation interior is

approximately 600m, compared to about 1.5 km at Liparoo. It is possible that the

smaller plantation patch size at Wemen contributed to the increased abundances

and enhanced community structure I saw here; however, more detailed sampling

would be necessary to isolate the effects of ‘edge’ and ‘area’ in this system

(Fletcher et al. 2007).

5.4.4 Conclusions

In my study system, I found that the woodland/plantation boundary was not a

barrier to wild pollinators, and spatiotemporal changes in resources and vegetation

heterogeneity across the ecotone were more likely influencing pollinator

distribution patterns and movement into the interior of the plantation. In

particular, I found that pollinator community similarity between sites was

93

associated with the relative level of heterogeneity between vegetation types, and

that each insect group's response to the edge and overall distribution patterns

varied across time – before, during and after the mass-flowering event in almond

plantations. My results concur with studies of similar focus that have found

greater influence from context-specific environmental variables (e.g. light

availability, canopy or shrub cover) (Yu et al. 2007; Muff et al. 2009) than

distance from the edge alone, and this interaction may be more pertinent when

considering highly-mobile insects like pollinators (Haslett 2001).

The significance of my findings would be augmented by more intense

sampling, beyond the 300 m distance limit used here and across other

woodland/plantation boundaries, as well as further consideration of changes in

environmental variables relative to the edge and surrounding landscape.

Nonetheless, my results provide valuable information for the ecological

management of conventional almond plantations, as well as pollinator

conservation in agroecosystems generally, and also contribute to knowledge of

how edges influence beneficial insects in managed landscapes. I showed that the

spatial distribution and structure of potential almond pollinator communities can

vary in response to spatiotemporal changes in resources, which is critical

information for developing effective pollinator conservation initiatives in

agroecosystems centred on pollinator-dependent crops. Future studies of how

habitat boundaries in agroecosystems influence important insect taxa should

survey both sides of the boundary and consider the complex interactions of

physical landscape configuration, spatial and temporal variation in resource

availability, and the unique responses of specific insect groups to these elements.

94

95

Chapter 6: Food and habitat resources available to wild pollinators in a

winter-flowering crop

Abstract

Homogeneous tree crop plantations can adversely impact wild pollinator

communities by limiting temporal continuity of food and the availability of

nesting sites. Identifying structural differences between plantations and adjacent

natural vegetation, and how these differences influence pollinator communities,

can inform ecological management goals. Communities of potential wild

pollinators (native bees, wasps, flies) were compared between monoculture

almond plantations and native mallee vegetation in a semi-arid region of

temperate Australia, before, during and after the late winter almond flowering

period. I examined whether abundance and richness of each insect group were

related to overall site heterogeneity or individual structural attributes of vegetation

at each site, focusing on food and nesting resources. Relationships with site

heterogeneity varied temporally, and between taxa, before, during and after

almond flowering. Insect response variables were often influenced by particular

habitat attributes when they were not associated with overall heterogeneity. Fly

and wasp richness and wasp abundance were strongly associated with leaf litter

and ground cover vegetation before almond flowering, regardless of vegetation

type. After flowering ceased, native bee abundance had a strong negative

relationship with canopy cover and leaf litter. This study provides novel insights

into an understudied ecosystem and highlights the value in focusing on specific

resources when investigating relationships between insects and habitat structure,

rather than generalised ‘heterogeneity’ metrics. In particular, maintaining

keystone resources in agroecosystems can enhance the conservation of insects and

ecosystem services, especially in areas where agricultural systems overlap

dynamic natural ecosystems.

*This chapter has been published as Saunders ME, Luck GW, Gurr GG (in press)

Keystone resources available to wild pollinators in a winter-flowering tree crop

plantation. Agricultural and Forest Entomology.

96

6.1 Introduction

Heterogeneity in vegetation structure is essential for supporting biodiversity and

ecosystem functions at multiple scales (e.g. Srivastava 2006; Diacon-Bolli et al.

2012), especially in agricultural systems (Altieri 1999; Benton et al. 2003;

Letourneau et al. 2011). At small scales, variation in vegetation structure and

resource availability can influence population dynamics and distribution patterns

of invertebrates, including ecosystem service providers like pollinators (Williams

& Kremen 2007; Murray et al. 2012) and natural enemies (Landis et al. 2000;

Gámez-Virués et al. 2010). Maintaining pollination and pest control ecosystem

services is vital for minimising ‘yield gaps’ in agricultural production (Bommarco

et al. 2013). Yet, modern agroecosystems are dominated by large patches of

homogeneous vegetation (e.g. broadacre crop monocultures) that can disrupt the

provision of these services (Tscharntke et al. 2005) and negatively impact local

wildlife populations (Krebs et al. 1999; Tilman 1999; Stoate et al. 2009). The

provision of ecosystem services in agroecosystems depends on complex trophic

interactions between multiple individuals, species and communities (Vandermeer

et al. 2010; Lundin et al. 2013) and vegetation homogeneity can disrupt these

interactions in agroecosystems. A number of studies have found that conversion

of heterogeneous landscapes to monoculture crop systems has been linked to

increased pests and diseases and subsequent declines in crop yield (Marshall &

Foot 1979; Altieri & Nicholls 2002; Aragona & Orr 2011; Bennett et al. 2012).

Wild pollinators are particularly vulnerable to intensive management, as

they require access to multiple heterogeneous resources throughout the year and

prefer to forage in plant-diverse systems (Williams & Kremen 2007; Winfree et

al. 2011). Numerous studies have shown that pollinator richness and abundance

generally increases with greater plant diversity or structural complexity in

managed crop systems (Nicholls & Altieri 2013), but relationships with these

factors can vary between taxa (Steffan-Dewenter et al. 2002; Holzschuh et al.

2010; also Chapter 3) and across different spatial and temporal scales (Kleijn &

van Langevelde 2006; Vasseur et al. 2013). Temporal variation in resources is

particularly relevant to studies of pollination services, as many pollinator-

dependent crops grown in monocultures exhibit a brief, highly-concentrated

resource pulse during flowering, yet provide very few resources for pollinators for

97

the remainder of the year. This phenological pattern may provide temporary

benefits for pollinator insects (Bartomeus & Winfree 2011; Lander et al. 2011),

but can have longer-term negative consequences for nearby native plant

pollination (Holzschuh et al. 2011) and insect reproduction (Westphal et al. 2009),

especially for species that emerge after the crop has ceased flowering (Jauker et

al. 2012).

Perennial tree crop plantations (e.g. orchard fruits) are an interesting study

system, as they are usually managed as monocultures across large areas, but are

more permanent in the landscape compared to annual grain or vegetable crop

systems (Hartemink 2005). Because of this, tree crops may have underestimated

potential for wildlife conservation, as long as production goals are balanced by

ecological management objectives (Hartley 2002; Barlow et al. 2007; Quine &

Humphrey 2010; Tscheulin et al. 2011). Many economically-valuable tree crops

are also pollinator-dependent (Cunningham et al. 2002; Losey & Vaughan 2006;

Calderone 2012), so knowledge of relationships between wild pollinators and

vegetation structure within these systems will aid in prioritising conservation and

management needs. The ability of tree crop plantations to support wild pollinator

communities is particularly important for pollinator-dependent orchard fruits that

flower in late winter, as most orchard fruits do. Pollinator activity can be limited

at this time, affecting fruit set of pollinator-dependent crops (Stoddard & Bond

1987; Bosch & Kemp 2002), yet few studies have investigated the capacity of

these systems to provide food or nesting resources that support overwintering and

emerging wild pollinator species.

Here, I sampled potential wild pollinator insects in monoculture almond

plantations and adjacent native vegetation in northwest Victoria, Australia, during

the almond flowering period. Almond (Prunus dulcis, Mill.) is the biggest

contributor to Australia’s total annual nut production, and production is centred

around the semi-arid mallee regions of northwest Victoria and eastern South

Australia (Almond Board of Australia 2013), which include areas of high

conservation value (Mallee Catchment Management Authority 2008; Cheal et al.

2013). Almond flowers require cross-pollination to set fruit and, because trees

flower in late winter when insect activity is limited, growers rely on imported

managed honeybee hives to provide pollination services and information on

potential wild pollinators available in Australia’s growing regions is limited. In

98

general, research on invertebrate ecology in the Victorian mallee is dominated by

studies of ants and information on the ecology of other species is limited;

however, the region is considered to have high invertebrate species richness,

particularly hymenopterans, and high phenological diversity in invertebrate

activity patterns, sometimes across very short time periods (Yen et al. 2006). In

addition, knowledge from other Mediterranean regions suggests that insect

pollinators could be active throughout the year (Michener 1979; Petanidou &

Lamborn 2008) and there is a need for studies that identify how the expansion of

tree crop plantations in the Australian mallee regions might influence local

pollinator communities.

I refer to individuals in the taxonomic orders of Diptera and Hymenoptera

as potential pollinators, hereafter simply pollinators, although I did not quantify

almond pollination. Fly and wasp species have been documented as pollinators of

native flowers in the study region (e.g. Ford & Forde 1976; Horskins & Turner

1999) and I observed many of these individuals visiting almond flowers during

sampling. In addition, Diptera and non-bee Hymenoptera species are important,

but understudied, generalist pollinators in a variety of natural and agricultural

systems (Larson et al. 2001; Rader et al. 2013), especially in Australia (Armstrong

1979; Pickering & Stock 2003), and many dipteran and hymenopteran species that

are generally categorised as ‘natural enemies’ are also common, but

unacknowledged, flower visitors and pollinators (Larson et al., 2001; Wäckers et

al., 2005). I investigated the relationships between these insect groups and site

attributes relevant to vegetation structure and resource factors, before, during and

after almond flowering. I also explored associations between September

abundance counts of herbivore insect groups collected at each site (Psocoptera;

Hemiptera: Aphididae, Cicadellidae) and abundance of flies and wasps in each

vegetation type, as these insects are common host or prey species for many flies

and wasps and are therefore another potential resource factor in my study system.

In particular, I considered whether monoculture almond plantations provide

sufficient food and habitat resources for overwintering and emerging pollinator

species by asking the following questions: (i) How do vegetation structure and

resource attributes differ between mallee woodlands and almond plantations,

throughout the flowering period? (ii) Does the relationship between site

heterogeneity and abundance or richness of pollinator insect groups vary across

99

the almond flowering period? (iii) How is the abundance and richness of

pollinators related to the availability of understorey attributes and resources?

6.2 Methods

6.2.1 Study area and vegetation structure

I sampled 68 sites across two conventional monoculture almond plantations (n =

50) and adjacent mallee woodlands (n = 18) in the Victorian Murray Mallee

Bioregion, northwest Victoria, Australia (-34.8 °, 142.6°). Average annual rainfall

in the study area is 314.9 mm (Station 076000; Bureau of Meteorology 2013). I

sampled on three occasions in 2011, before (July), during (August) and after

(September) the mass-flowering event in almond plantations, which typically lasts

3-4 weeks. Plantations that were adjacent to large patches of remnant mallee

vegetation were chosen for sampling (see Appendix 5 Table A5.1). Liparoo

plantation is adjacent to the Annuello State Flora and Fauna Reserve and Wemen

plantation is adjacent to an unnamed patch of remnant woodland. Both plantations

are broadacre monocultures, intensively managed by the same company and the

linear distance between the two plantations is approximately 7.5 km. Plantations

had identical row widths (5-6 m) and tree spacing (5 m) and herbicides were used

to maintain bare ground across plantation floors.

Almond trees are deciduous and plantations were completely bare of

leaves or flowers before flowering (July). In August, mass-flowering occurred

across the entire plantation, while each tree was still predominantly leafless

(approx. 50,000 individual blossoms per tree). In September, the majority of

almond blossoms had dropped and trees were fully leaved. Mallee woodlands are

open, shrubby evergreen woodlands on alkaline, semi-arid soils and are restricted

to Mediterranean climate zones of southern Australia. They have heterogeneous

middle- and upper-storey vegetation, dominated by Eucalyptus species growing in

the characteristic ‘mallee’ habit of shorter trees with multiple trunks, as well as

Acacia, Callitris and Casuarina tree and shrub species. The understorey is highly

variable, with areas of bare ground interspersed with patchy ground cover

typically including hummocks of Triodia, halophytic plants, herbs and grasses,

and dense mats of leaf litter near eucalypt trees. Phenological changes in

100

woodland structure were less obvious across the sampling period, compared to

plantations, with evergreen vegetation and diverse, patchy floral resources

available in all months (see Figure 5.2 for representative photographs).

6.2.2 Data Collection

Flying insects were collected using pan traps and data on local habitat attributes

were recorded at each site. Pan trapping was chosen over on-site visual

identification methods, as it does not suffer from the ‘collector bias’ inherent in

the latter (Westphal et al., 2008); however, colour and placement of traps was

designed relevant to the context of the system and research questions, to increase

the chance that the data collected were ecologically meaningful (see Saunders &

Luck 2013; Appendix 1). One white and one yellow pan trap were placed at each

site, as white was the predominant flower colour in plantations and yellow the

most common in woodlands. Pan traps were placed on the ground under flowering

trees for maximum visibility, and because I was interested in associations between

ground-level resources (leaf litter, ground cover vegetation) and pollinator insects

that use these resources for foraging and nesting sites. As discussed in section

6.2.1, mats of leaf litter and herbaceous/weedy ground cover are the most

common ground-level resources in both vegetation types, and these elements are

eseential nesting and foraging habitat for flies, non-Apis bees and wasps (Leather

et al. 1993; Austin et al. 2005; Michener 2007). Traps were open for 8 hours on

fine, calm days, and were set just after sunrise and collected just before nightfall

in the order they were placed. Insects were transported to the laboratory in 70%

ethanol, counted and sorted into taxonomic groups (Diptera; Hymenoptera (non-

Apis bees and wasps); Hemiptera (Aphididae, Cicadellidae); Psocoptera). I

identified native bees to species (where possible, although most Lasioglossum

individuals could only be classified to subgenus) and wasps and flies to family

groups using identification keys (Hamilton et al. 2006; Stevens et al. 2007) along

with a reference collection provided by a taxonomic expert (K. Walker, Museum

Victoria).

At each trap site, I collected data on vegetation structure, as well as

canopy cover in September, as almond plantations were bare of leaves during July

and August. I did not quantify floral diversity or abundance in this study, because

101

plantations contained a single plant species (almond) that flowered in August only

and no flowering weedy plants were present under almond trees, so statistical

comparisons with woodland floral diversity across months would not be possible.

Small-scale vegetation structure was quantified using a line intercept method. At

each trap site, a 10 m line transect was marked out in a randomly-chosen direction

and the proportion of the transect length covered by the following structural

attributes was recorded: bare ground; leaf litter or decaying wood/nuts;

herbaceous ground cover vegetation (forb, succulent, moss); annual/perennial

grass; Triodia hummock grass; shrub <1 m; tree 1—2.9 m; tree 3—4.9 m; tree >5

m. In September, canopy cover was measured by placing a digital camera facing

upwards on the ground beside each trap and photographing the canopy above the

trap site. Images were then converted to black and white and the percentage of the

image area that was covered with canopy vegetation was quantified using

ImageTool 3.0, a freeware image processing and analysis program (UTHSCSA

2002, available from URL http://compdent.uthscsa.edu/dig/itdesc.html).

6.2.3 Data Analysis

Analyses of pollinator groups were conducted on abundance counts and a richness

index. Richness was calculated using the Menhinick index (Menhinick 1964).

Insect abundance counts were not transformed, as this approach is not

recommended when analysing overdispersed ecological count data (O’Hara &

Kotze 2010; Warton & Hui 2011). Raw data from the vegetation structure

transects were used to calculate a heterogeneity score for each site using the

complement of Simpson's Index (1-Simpson's Dominance measure): a score of 1

means that all habitat attributes were equally present along the measured transect

(high heterogeneity) and a score of 0 means one habitat attribute dominated the

entire transect (low heterogeneity).

Differences between the two vegetation types were identified for each

habitat attribute and insect group using Mann-Whitney (U) pairwise tests, and

then pollinator community composition was compared between woodlands and

plantations using ordination and an associated permutation test for significance. I

used non-metric multidimensional scaling (NMDS) based on Bray-Curtis

similarities between sites to illustrate overall pollinator community composition

102

for the whole sampling period (pooled monthly abundance counts of each

pollinator insect group). The 'goodness-of-fit' for the configuration was confirmed

with an accompanying Shepard plot and STRESS value (values < 0.1 indicate that

the plot shows an ideal configuration). To identify whether the pattern shown in

the NMDS ordination was representative of the whole sampling period, a one-way

ANOSIM (analysis of similarities; Clarke 1993) was used to test for significant

differences in monthly pollinator communities grouped by the two vegetation

types, mallee or almond. The analysis provides an R statistic between -1 and +1:

values < 0.25 indicate that similarity coefficients within the compared groups are

similar to coefficients between groups, so the groups cannot be separated; values

> 0.50 indicate that the groups can be separated, but there is some overlap; and

values > 0.75 indicate distinct separation between groups (Chapman &

Underwood 1999).

6.2.4 Relationships between potential pollinators and habitat attributes

Relationships between pollinator groups and habitat attributes were identified

using generalised linear models (GLMs). First, I examined the relationship

between site heterogeneity scores and monthly richness and abundance of each

insect group, and compared the fit of each insect/heterogeneity model with the

intercept-only model. I then examined relationships between insect richness and

abundance and the understorey attributes that were present in both vegetation

types, ‘leaf litter’ and ‘ground cover’. I focused on these attributes as they were

the most biologically meaningful to my study questions, because many of the

target insect species nest or overwinter in the ground, leaf litter or ground cover

vegetation (Leather et al. 1993). ‘Canopy cover’ was also included in models of

September insect assemblages, as this attribute affects understorey light

availability and can influence insect abundance (Michener 2007; Abrahamczyk et

al. 2010).

GLMs were based on combinations of selected predictors. Monthly insect

abundances were modelled with a negative binomial distribution and richness data

were modelled with a linear distribution. The number of parameters (inclusive of

the intercept) in each model was limited to a maximum of six, to avoid over-

fitting. I included vegetation type (mallee, almond) as a categorical factor, and

103

percent ground cover, leaf litter (all months) and canopy cover (September models

only) as covariates. Interaction terms between vegetation type and each attribute

were included, as I believed that the importance of ground cover, leaf litter and

canopy cover may differ in plantations compared to mallee woodlands. I included

the intercept-only model for all response variables as a baseline comparison.

Competing explanatory models for each response were compared using an

information-theoretic model selection process. I compared information criterion

values within each set of models to rank models based on goodness-of-fit and

complexity. Richness models were compared by the modified Akaike’s

Information Criterion for small samples (AICc) and abundance models by the

quasi-likelihood adjusted AICc (QAICc) for overdispersed count data. Hereafter,

both AICc and QAICc values are referred to as simply AICc. Differences in AICc

(Δi) were calculated relative to the smallest AICc value in each model set and a

candidate model set for each response (all models with Δi < 10) was examined.

Most of the relationships were explained by the Δi < 2 model set, so only these are

presented here (see Appendix 5 Table A5.2 for full list of candidate models). I

calculated relative log likelihoods, L(gi) = exp(-0.5*Δi), and AICc weights, wi, for

each model to assess the probability of each model being a better fit to the data

relative to other models in the set and summed wi over all models in the set

containing that predictor to obtain an estimate of the relative importance of each

explanatory variable. In September model sets, the ‘Veg Type’ variable was

included in only five models, while all other variables were included in six

models. Therefore, all September variable weights were calculated from only the

five highest-ranked models for each variable, to balance the number of models

containing each variable.

To explore associations between Hymenoptera and Diptera and

herbivorous insect groups I used association analysis as an exploratory tool for

future hypothesis-generation, as I did not identify herbivores to species, nor did I

measure trophic interactions between parasitoids and prey species in my system.

To minimise the potential distortion of species associations arising from a high

number of zeros in the abundance data, I converted abundances for all five insect

groups to presence/absence data. I used Sørensen’s coefficient (Ssor) for

presence/absence data, as my focus was on co-occurrence of species per site

(Legendre & Legendre 1998). All analyses involving native bees and herbivores

104

were conducted on September data only, as these insect groups were mostly

absent in July and August. GLM analyses were conducted in IBM SPSS 20 (IBM

Corporation 2011) and I used PAST 2.17 (Hammer et al., 2001) for all other

analyses.

6.3 Results

6.3.1 General observations and overall community composition

I caught 2476 flies in 20 family groups, 336 wasps in 23 families, and 171 native

bees in 2 genera (see Appendix 2 Table A2.3). Fly abundance was significantly

higher in the plantations in all months, but monthly richness was not significantly

different between vegetation types. Wasp abundance and richness in July were

significantly higher in mallee, but wasp and bee abundance and richness in other

months were not significantly different between vegetation types (Table 6.1). The

cover of canopy and bare ground were higher at almond sites, while overall

habitat heterogeneity and all other vegetation attributes were higher at mallee sites

(Table 6.1). Cicadellidae and Psocoptera abundance were higher in plantations

than in woodlands, but Aphididae abundance was not different between

vegetation types (Table 6.1).

Overall pollinator community composition across the sampling period was

different between vegetation types, although there was some overlap between

sites (Figure 6.1). Results of the ANOSIM test showed that mallee and almond

pollinator communities differed significantly in all months, but the most variation

in similarity coefficients occurred in September (R = 0.44, p < 0.001). Community

similarity between vegetation types was only weakly separated in July (R = 0.15,

p = 0.02) and August (R = 0.16, p = 0.01).

105

Table 6.1 Mean (±SD) monthly insect abundance (A) and richness (R) of each pollinator group and mean (±SD) values for all resource attributes per trap site in mallee vegetation and almond plantations.

Mallee (n = 18) Almond (n = 50)

Pollinator group

Flies Jul (A) 7.2 (5.4)*

(R) 1.7 (0.6)

(A) 3.5 (2.5)

(R) 1.5 (0.7)

Aug (A) 10.3 (7.3)

(R) 2.1 (0.7)

(A) 17.1 (11.5)*

(R) 2.1 (0.6)

Sep (A) 7.4 (5.5)

(R) 1.9 (0.5)

(A) 19.9 (10.8)*

(R) 1.9 (0.6)

Wasps Jul (A) 1.6 (1.4)*

(R) 1.0 (0.6)*

(A) 0.6 (1.2)

(R) 0.4 (0.6)

Aug (A) 2.4 (4.3)

(R) 0.7 (0.6)

(A) 0.9 (1.0)

(R) 0.6 (0.6)

Sep (A) 3.5 (3.9)

(R) 1.3 (0.5)

(A) 2.5 (2.4)

(R) 1.2 (0.8)

Native bees Sep (A) 4.5 (9.8)

(R) 0.6 (0.5)

(A) 1.7 (2.7)

(R) 0.5 (0.5)

Resource attributes

Overall habitat heterogeneity 0.73 (0.2)* 0.19 (0.1)

Bare ground 0.15 (0.1) 0.89 (0.1)*

Leaf litter/decaying wood 0.21 (0.1)* 0.04 (0.04)

Ground cover vegetation 0.1 (0.1)* 0.01 (0.03)

Shrub < 1 m 0.18 (0.3) 0

Grasses 0.06 (0.07)* 0.007 (0.05)

Triodia hummock 0.03 (0.07) 0

Tree 1-3 m 0.1 (0.1) 0

Tree 3-5 m 0.09 (0.1)* 0.01 (0.02)

Tree > 5 m 0.09 (0.1) 0.04 (0.03)

Canopy cover^ 0.26 (0.1) 0.58 (0.2)*

106

Herbivorous insects

Cicadellidae^ (A) 0.56 (0.8) (A) 3.28 (3.2)*

Psocoptera^ (A) 0.22 (0.4) (A) 1.92 (2.7)*

Aphididae^ (A) 0.22 (0.6) (A) 0.48 (1.1)

^ data are from September only; Asterisks indicate in which vegetation type each value was significantly greater, * p < 0.01. Overall habitat heterogeneity = Simpson’s index, 1-D; Vegetation resource attributes = proportion of cover along 10 m transect

Figure 6.1 Non-metric multidimensional scaling ordination of overall pollinator community similarity (Bray-Curtis) between mallee woodlands and almond plantations. The configuration shown is an ideal representation of the data (STRESS value = 0.07).

6.3.2 Site heterogeneity

The influence of overall habitat heterogeneity at each site on pollinator groups

varied across the sampling period (Table 6.2). Fly abundance was higher at more

heterogeneous sites in July, but was associated with increased homogeneity in

August and September. Fly richness in all months was not influenced by habitat

heterogeneity. Wasp abundance in July and August, as well as richness in July,

Mallee

Almond

107

were associated with higher habitat heterogeneity. The following groups were not

associated with site heterogeneity: August wasp richness; September wasp

richness and abundance; native bee richness and abundance.

6.3.3 Understorey attributes and resource factors

The responses of each insect group to understorey attributes varied across the

flowering season (Tables 6.3 & 6.4). In July, fly abundance showed the strongest

association with ground cover vegetation, which had a summed Akaike weight of

0.84 (Table 6.4). However, the parameter estimates for ‘ground cover’ in the two

highest-ranked models had broad confidence intervals encompassing zero and the

vegetation type/ground cover interaction term in both models showed that fly

responses to ground cover differed between vegetation types (Table 6.3). Pearson

correlations between abundance and ground cover for each vegetation type

confirmed that the presence of ground cover was positively correlated with fly

abundance in mallee (r = 0.43, p = 0.07), but was not associated with abundance

in almond plantations (r = -0.17, p = 0.24). The third-highest ranked model

suggested a positive relationship between leaf litter and fly abundance in July,

regardless of vegetation type, but the likelihood of this model was low (L(gi) =

0.39) and the relative importance of leaf litter was lower than either ground cover

or vegetation type (wi = 0.45; Tables 6.3 & 6.4).

In August, the model containing only vegetation type as a predictor

provided the best fit for fly abundance data (Table 6.3), but leaf litter had the

highest relative influence compared to the other predictors (Table 6.4) and this

attribute had a strong negative relationship with fly abundance (Table 6.3). Leaf

litter and ground cover were associated with fly abundance in September, but

vegetation type had the highest relative influence (wi = 1) (Table 6.4). Vegetation

type also affected the fly—ground cover relationship and correlation analysis

showed that fly abundance was positively associated with ground cover vegetation

in almond plantations (r = 0.35, p = 0.01) but was not associated with ground

cover in mallee woodlands (r = -0.22, p = 0.38). Fly richness showed a weak

positive association with leaf litter in July and ground cover in August (Tables 6.3

& 6.4), but the intercept-only model was ranked as the best fit to the fly richness

data in both months. September fly richness was negatively associated with

108

Table 6.2 Relationships between overall site heterogeneity and abundance and richness of each pollinator insect group in each month.

Pollinator group MonthAbundance Richness

Model χ2 wi Model χ2 wiFly July 0.59 (0.17) 18.2** 0.99 0.35 (0.28) 1.6 0.43

August -12.36 (3.60) 8.5** 0.93 -0.26 (0.28) 0.9 0.34

September -19.39 (3.50) 18.9** 0.99 -0.15 (0.25) 0.4 0.29

Wasp July 1.30 (0.44) 11.6** 0.92 1.16 (0.24) 19.9** 0.99

August 0.52 (0.24) 5.1* 0.78 -0.05 (0.28) 0.03 0.25

September 0.03 (0.16) 0.02 0.26 -0.21 (0.30) 0.5 0.30

Bee September -2.13 (1.41) 1.7 0.35 0.16 (0.20) 0.70 0.32

* p < 0.05; ** p < 0.01 Model: parameter estimates with SE in parentheses; χ2: likelihood ratio test statistic, indicates if the model provides a significantly better fit for the data set than the ‘constant only’ model; wi: Akaike weight for the heterogeneity model, gives the probability that the model is the best fit for the relationship

109

Table 6.3 Relationships between resource attributes and monthly pollinator insect abundance (A) or richness (R) across all almond and mallee sites (n = 68). The highest-ranked models (Δi AICc ≤ 2) in each candidate set are shown. Where the constant only model is ranked the highest, only the second-highest model is listed for comparison. All models include the Intercept. See Appendix 5 Table A5.2 for the full candidate set of models for each response variable. Response Month Model Δi AICc L(gi) wiFly (A) July Veg (0.47; 1.20) + GC (-15.65; 9.76) + Veg*GC (49.04; 17.14)

Veg (1.64; 2.17) + GC (-13.81; 9.72) + LL (16.78; 9.62) + Veg*GC (49.19;

18.86) + Veg*LL (-19.6; 13.95)

GC (15.5; 8.35) + LL (11.23; 5.60)

0

1.36

1.87

1

0.51

0.39

0.41

0.21

0.16

August Veg (-6.78; 2.19)

LL (-28.99; 8.20)

Veg (0.63; 4.29) + LL (72.76; 50.13) + Veg*LL (-94.53; 51.81)

GC (-17.36; 16.54) + LL (-25.65; 9.22)

0

0.31

0.46

1.81

1

0.86

0.80

0.41

0.28

0.24

0.22

0.11

110

September Veg (-0.76; 0.21) + GC (5.56; 2.60) + Veg*GC (-7.37; 3.0)

Veg (-1.32; 0.30) + LL (3.07; 1.91) + Veg*LL (-1.06; 2.21)

Veg (-0.99; 0.16)

Veg (-1.54; 0.33) + CC (0.12; 0.48) + GC (1.39; 1.34) + LL (2.63; 0.99)

0

0.38

1.34

1.84

1

0.83

0.51

0.40

0.35

0.29

0.18

0.14

(R) July LL (1.27; 0.71)

Con (1.50; 0.08)

GC (0.69; 1.14) + LL (1.18; 0.72)

0

0.96

1.26

1

0.62

0.53

0.30

0.19

0.16

August Con (2.12; 0.08)

Veg (-0.13; 0.22) + GC (-7.84; 3.18) + Veg*GC (7.96; 3.50)

0

0.88

1

0.65

0.29

0.19

September Veg (-0.8; 0.03) + CC (-1.05; 0.45) + GC (0.19; 1.22) + LL (2.06; 0.91)

Veg (-0.63; 0.27) + LL (0.12; 1.94) + Veg*LL (2.47; 2.18)

Con (1.92; 0.07)

0

1.54

1.58

1

0.46

0.45

0.26

0.12

0.12

111

Wasp (A) July LL (4.45; 2.03)

Veg (0.96; 0.43)

GC (2.57; 2.67) + LL (3.90; 1.94)

0

0.04

1.43

1

0.98

0.49

0.30

0.29

0.15

August GC (1.95; 0.68)

Veg (0.71; 0.26) + LL (0.26; 2.93) + Veg*LL (-1.89; 3.07)

Veg (0.41; 0.16)

0

0.15

0.37

1

0.93

0.83

0.28

0.26

0.23

September Con (-0.49; 0.12)

Veg (0.38; 0.16) + LL (0.54; 1.49) + Veg*LL (-1.95; 1.66)

0

0.63

1

0.73

0.19

0.14

(R) July GC (2.36; 1.04) + LL (1.93; 0.66) 0 1 0.53

August Con (0.64; 0.08)

GC (0.79; 1.13)

0

1.71

1

0.43

0.41

0.18

September Con (1.20; 0.08)

GC (-1.19; 1.23)

0

1.27

1

0.53

0.27

0.14

112

Native Bee (A) September CC (-4.83; 0.88) + LL (-4.05; 1.79) + GC (-5.45; 2.74)

Veg (0.78; 0.63) + CC (-3.89; 1.08) + GC (-6.61; 2.85) + LL (-5.58; 2.14)

LL (-4.72; 1.84) + CC (-4.74; 0.91)

0

0.94

0.97

1

0.63

0.62

0.37

0.23

0.23

(R) September Con (0.50; 0.06)

GC (1.02; 0.81)

0

0.62

1

0.73

0.21

0.15

Veg = vegetation type; CC = canopy cover; LL = leaf litter; GC = ground cover; Con = constant only model. Vegetation type is a categorical factor, testing whether pollinator-habitat relationships differed between the two vegetation types. Model: parameter estimates and SE for each predictor in parentheses; * indicates an interaction term; Δi AICc = difference in AICc values, relative to the highest-ranked model; L(gi) = relative likelihood of the model; wi = Akaike weight for the model.

113

Table 6.4 Summed Akaike weights (wi) for each resource attribute included as an explanatory variable in generalised linear models, indicating the relative importance of each attribute for explaining variation in the response variables (insect abundance or richness). Response Month Canopy Cover Ground Cover Leaf Litter Vegetation Type

Fly (abundance) July 0.84 0.45 0.75

August 0.22 0.60 0.56

September 0.17 0.49 0.43 1

(richness) July 0.35 0.52 0.25

August 0.45 0.27 0.40

September 0.50 0.47 0.53 0.50

Wasp (abundance) July 0.27 0.50 0.38

August 0.50 0.39 0.61

September 0.25 0.27 0.36 0.34

(richness) July 0.65 0.78 0.31

August 0.26 0.23 0.22

September 0.23 0.36 0.23 0.25

Bee (abundance) September 0.99 0.72 0.82 0.24

(richness) September 0.30 0.34 0.25 0.22

Canopy cover was not included in July and August models. All September variable weights were calculated from the five highest-ranked models for each variable, to balance the number of models used to calculate wi

114

canopy cover and positively associated with leaf litter (Table 6.3), but all four

predictors had relatively equal importance (Table 6.4).

Wasp richness in July showed the strongest (positive) relationship with

both ground cover and leaf litter and these relationships were not affected by

vegetation type (Tables 6.3 & 6.4). Wasp abundance in July and August showed

the strongest (positive) relationship with leaf litter and ground cover respectively,

regardless of vegetation type, although the model containing vegetation type alone

was also ranked highly in both months (Table 6.3). None of the predictors were

related to wasp richness in August and September or abundance in September

(Tables 6.3 & 6.4). September native bee abundance had a strong negative

relationship with canopy cover, ground cover and leaf litter, but was not

influenced by vegetation type (Table 6.3). Canopy cover had the strongest relative

influence on bee abundance compared to the other attributes (wi = 0.99; Table

6.4). Bee richness showed a weak association with ground cover (Tables 6.3 &

6.4).

At mallee sites, fly and wasp presence showed a moderate association with

Cicadellidae presence (wasp, Ssor = 0.58; fly, Ssor = 0.56), but co-occurrence with

Psocoptera (wasp, Ssor = 0.38; fly, Ssor = 0.36) and Aphididae (wasp, Ssor = 0.30;

fly, Ssor = 0.29) was much weaker. In almond plantations, wasps and flies were

strongly associated with Cicadellidae (wasp, Ssor = 0.85; fly, Ssor = 0.91) and

Psocoptera presence (wasp, Ssor = 0.69; fly, Ssor = 0.70), but Aphididae presence

was not strongly associated with either wasps (Ssor = 0.39) or flies (Ssor = 0.39).

6.4 Discussion

6.4.1 Site heterogeneity

Although the composition of potential wild pollinator assemblages differed

significantly between mallee woodlands and almond plantations in each month,

there was a lot of overlap in assemblages for the whole sampling period,

suggesting that pollinator community composition in mallee or almond vegetation

was more likely associated with variation in smaller-scale environmental factors

rather than the overall vegetation type. The strongest compositional difference

between mallee and almond assemblages was in September, possibly influenced

115

by the significant increase in the number of insects collected in plantations in that

month. Some of the increased abundance is likely accounted for by warmer spring

temperatures which would have prompted the emergence of more adult insects,

especially bees (average daily min-max temperature range for study area: 4.3 °C –

15.1 °C in July compared to 6.7 °C – 20.2 °C in September; Bureau of

Meteorology 2013). The relationship between each pollinator group and site

heterogeneity also varied across months. Abundance and richness of flies and

wasps were associated with higher heterogeneity in July, and wasps were also

more abundant at heterogeneous sites in August, during the almond flowering

event. However, fly abundance was negatively related to site heterogeneity in

August, as well as in September, after flowering had finished, when they were the

only pollinator group to show a relationship with site heterogeneity.

This variation in insect-site heterogeneity associations across such a short

time period reinforces the idea that the relationship between farmland biodiversity

and agricultural intensification is not a simple linear one (Burel et al. 1998) and an

insect’s response to agricultural environments can be more influenced by its life

stage (e.g. Delettre 2005) or seasonal environmental changes (e.g. Reese &

Philpott 2010) than on simple metrics of overall habitat heterogeneity. Tews et al.

(2004) highlight the importance of considering how ‘heterogeneity’ metrics are

measured, as the attributes used to calculate the metric may create beneficial

‘heterogeneity’ or detrimental ‘fragmentation’ for different taxa, depending on the

context. For example, my habitat heterogeneity score was based on the number of

different surface elements across a given stretch of land. Consequently, an area of

completely bare ground would receive a similar heterogeneity score as an area

completely covered by leaf litter, yet the latter would provide structural

heterogeneity and diverse micro-habitats and climates for a variety of taxa. Hence,

it would be reasonable to consider relationships between organisms and specific

structural attributes. The regression models for individual habitat attributes

corroborated this, as abundance or richness of some insect groups were strongly

related to individual attributes even when they were not related to overall site

heterogeneity (Tables 6.2 & 6.3).

116

6.4.2 Understorey attributes: winter

Many Diptera and Hymenoptera species are dependent on the thermal protection

provided by the litter layer or ground cover vegetation during winter, as they

either overwinter or nest there themselves, or prey on other species that live in

these microhabitats (Leather et al. 1993; Landis et al. 2000; Austin et al. 2005;

Michener 2007; Bale & Hayward 2010). This was reflected in my results, as a

high proportion of leaf litter and ground cover was particularly important for flies

and wasps in July, when almond plantations were devoid of resources in the upper

vegetation layer. Wasp abundance and richness were associated with leaf litter

and ground cover regardless of vegetation type, but the relationship between fly

abundance and ground cover was strongest in mallee woodlands, where the

proportion of ground cover vegetation was higher than in plantations. This

suggests that there may have been other resources or environmental factors

specific to mallee sites that influenced fly assemblages more than wasps. Wasp

richness in July was the only response variable to show the strongest relationship

with both ground cover and leaf litter together, but leaf litter had the greatest

relative importance. There is a paucity of knowledge on the ecology of wasp

species in my study system specifically; however, leaf litter provides vital habitat

and foraging substrate for various Australian wasp families, including scelionid

species (Austin et al. 2005), which were common in samples in July, and litter

patches are known to be essential habitat for other hymenopteran taxa, i.e. ants, in

the Victorian mallee (Andersen 1983; Yen et al. 2006). Although some of my

sampled wasps were likely parasitoid species, many carnivorous wasps are also

nectar-feeders and their effectiveness as pollinators is often underestimated

(Wäckers et al. 2005). In addition, pollination and pest control services in

agroecosystems are synergistically linked (Bos et al. 2007; Lundin et al. 2013)

and more detailed investigation of this study system would contribute valuable

information to our knowledge of these interactions. I was unable to determine if

understorey attributes were important for bees in this study, as I collected no bees

in July and only four individuals in August; however, this was most likely a result

of the passive sampling method, which was unable to detect hibernating or larval-

stage insects. Many of the bee species I collected are known to nest in the ground

(Knerer & Schwarz 1976; Houston & Maynard 2012) and active sampling for nest

117

sites and brood cells would reveal more information about their nesting habits in

my system.

6.4.3 Understorey attributes: late winter-spring

The relationships between wild pollinators and ground-level attributes in August

and September were weaker than in July, suggesting that the importance of these

attributes declined as the weather warmed and resources increased in both

vegetation types. There was no evidence of a convincing relationship between

August fly richness and any predictor, but the candidate set of models for August

fly abundance revealed a strong negative relationship with leaf litter. However,

vegetation type was ranked nearly as highly as leaf litter for explaining fly

abundance and, given that fly relationships with leaf litter in other months were

either positive or neutral, this most likely reflected that abundance was extremely

skewed toward plantation sites in August. The negative relationship with leaf litter

could be reflecting a positive relationship with ‘bare ground’, the most dominant

attribute in plantations (leaf litter was negatively correlated with the proportion of

bare ground, r = -0.73, p < 0.001). However, it is also possible that the negative

relationship indicated fly species dependent on leaf litter habitats were using the

plantation litter habitats more than mallee litter habitats. This could be because of

the additional resource supply available from nearby almond flowers (Fayt et al.

2006), or because of the increased connectivity of litter patches in plantations,

which were configured in continuous strips of litter along rows of almond trees

(Schiegg 2000; Randlkofer et al. 2010b). In mallee woodlands, areas of litter were

significantly larger, but more patchy, with areas of bare ground or vegetation

between litter patches, and this raises interesting questions for investigations of

potential pollinator insects using litter habitats in my system.

Wasps showed the strongest association with ground cover vegetation in

August, which had greater cover at mallee sites. Although I did not record plant

species at each site, many of the native understorey plants were flowering at

mallee sites during sampling and it is possible that these flowers were preferred

by foraging wasp species over almond flowers. There are many examples of local

adaptation and co-evolution between plants and their insect pollinators (Fægri &

van der Pijl 1979; Kalske et al. 2012), particularly among Australian

118

hymenopterans (Adams & Lawson 1993; Peakall & Handel 1993; Dyer et al.

2012). Mallee ecosystems are typical of Mediterranean bioregions in other parts

of the world, with a high diversity of annual and perennial herbs (Naveh &

Whittaker 1979) and flowering tree canopies close to the ground (Andersen &

Yen 1992), and wasps are common visitors to flowering herbs in European

Mediterranean ecosystems (Bosch et al. 1997). Hence, non-bee hymenopterans

native to this region may have co-evolved with, and therefore prefer, these

flowering plant species (Kummerow 1983) and this is an area with scope for

further research.

In September, wasps were not associated with any of the measured habitat

attributes, but wasp presence was related to the presence of Cicadellidae and

Psocoptera in plantations, as was fly presence. This result could indicate

parasitoid-host or predator-prey relationships, or simply a mutual preference for

similar habitat attributes, but further sampling and identification of potential

interactions between these insect groups is necessary to confirm the nature of this

effect. For example, fly abundance and richness were positively associated with

leaf litter (Table 6.3), as was Psocoptera abundance (r = 0.38, p = 0.007), which

could partly explain the fly-Psocoptera relationship I found. Fly abundance

showed a positive relationship with ground cover in September, but only in

plantations, and this result was most likely influenced by differences between the

two plantations I sampled. Ground cover was more common at Wemen plantation,

which has been established longer than Liparoo and contained predominantly

mature almond trees (approximately 25 years old). Stands of mature trees are

critical habitat elements for forest/woodland insect species, as they provide

established microclimates and habitat structures (e.g. dead wood, bark crevices,

leaf litter banks) that support diverse communities and complex food webs

(Warren & Key 1991). This link between invertebrate communities and mature

trees has also been found in orchard systems (Dubois et al. 2009; Herrmann et al.

2012) and predatory flies and wasps in particular may respond to the higher

resource availability in these environments.

Native bees showed the strongest response to vegetation structure in

September, compared to other pollinator groups. Bee abundance was negatively

associated with leaf litter, ground cover and canopy cover, suggesting a preference

for lighter, more open areas. It is possible that this result was influenced by two of

119

the mallee sites which were extreme outliers – 35 % of the total bee abundance

was collected at these two sites. Both sites had very low canopy cover and were

dominated by shrubs and heath vegetation. However, when I removed the two

outliers from the data set, overall bee abundance was still negatively correlated

with canopy cover (r = -0.34, p = 0.006) and this effect was present in plantations

(r = -0.65, p < 0.001), but not in woodlands (r = 0.12, p = 0.66). These results

agree with the majority of findings that suggest bees prefer open areas over dense

forests, particularly in rangeland, temperate or semi-arid systems (Grundel et al.

2010; Hoehn et al. 2010; Proctor et al. 2012). Leaf litter can be a limiting factor

for bee communities in Mediterranean environments because many species prefer

to excavate their nests into bare ground (Potts et al. 2005), including species in my

system (Knerer & Schwarz 1976; Houston & Maynard 2012), and this could

explain the negative relationship seen here. A similar effect could also explain the

negative relationship between bee abundance and ground cover. In contrast, I

found that bee richness was positively associated with ground cover vegetation,

although the relationship was very weak. This agrees with the strong wild bee-

ground cover relationships I found in Chapter 3. In addition, Yen et al. (2006)

observed that Halictine bees in Victorian mallee (which dominated my collected

bee fauna) appear to prefer understorey perennial flowering plants, particularly

around disturbed, weedy sites, and this is a promising area for further research

with the potential to inform management goals.

6.4.4 Conclusions and implications

This study has highlighted gaps in ecological knowledge of the species rich

mallee insect fauna. My findings would be augmented by further investigation of

the ecology and life history of pollinator insect species in the mallee, especially

around the rapidly expanding almond plantations. Further sampling across more

sites and over a longer time period, as well as more detailed taxonomic

identification and examination of species-habitat relationships, would enable

researchers to determine the contribution of these insects to ecosystem function in

mallee woodlands and adjacent almond plantations. In relation to this, my study

has provided two major insights. Firstly, I have shown that insects may be

strongly associated with particular habitat attributes, or keystone structures (Tews

120

et al. 2004), despite showing no relationship with overall heterogeneity. This

highlights the value in considering specific habitat and resource requirements of

the focal species when investigating relationships between ecosystem service

providers and habitat structure, rather than focusing on generalised

‘heterogeneity’ metrics. Secondly, although almond plantations supported a

diverse group of potential wild pollinators, the relationships between different

groups of pollinators and habitat attributes at each site varied among taxa, and

also with seasonal and phenological changes in the local environment. This

knowledge will be beneficial to plantation managers and may encourage the

adoption of holistic, ecologically-focused management strategies that provide

resource diversity within plantations to support multiple wild pollinator taxa, and

other ecosystem service providers, throughout the year. Hence, given the

conservation value of the mallee bioregions where Australian almond production

is concentrated and the gaps I have highlighted in our knowledge of habitat use by

mallee pollinator communities, this study system provides a unique opportunity

for further investigation of these issues and the potential to align conservation and

management goals.

121

PART III: SYNTHESIS

122

123

Chapter 7: Synthesis and Conclusions

7.1 Synthesis and general discussion

Ecological management of agroecosystems relies on understanding the biological

properties and ecological processes of the natural environment surrounding the

system (Pimentel et al. 1989). In particular, knowledge of local insect

communities and the way they interact with their environment can have

widespread benefits for biodiversity conservation and the productive capacity of

the agroecosystem. The work presented here has achieved two aims: it provides

insight into potential wild pollinator taxa living in the Victorian Murray Mallee

that could provide pollination and pest control services to almond growers; and it

gives credence to my overall thesis that intensively-managed monoculture tree

crop plantations do not provide the vegetation heterogeneity necessary to support

diverse native insect communities in the long-term. The data presented here were

limited in seasonal scope, being relevant to the almond flowering season in late

winter. The purpose of this was to provide information on the most critical life

stages for both the crop variety and the focal insect taxa. Almond trees rely

completely on insect pollination to set fruit, so the ability of a commercial almond

orchard to sustain a diverse pollinator community in the long-term is paramount.

This limitation is even more challenging for growers because almond trees flower

in winter when insect activity is limited. Additionally, given the natural dynamics

of semi-arid mallee ecosystems, where resource availability can be unreliable and

prone to extreme weather events, resource-rich agroecosystems could provide a

consistent resource supply for some insect species. Thus, prioritising ecological

management of these agroecosystems has potential to enhance the conservation of

both wild insects and the ecosystem services they provide.

My results have shown that potential wild pollinator taxa are using almond

plantations during flowering, as well as immediately before and after, but their

distribution throughout plantations appears to be limited by homogeneous

vegetation structure and lack of permanent resources. A number of results were of

particular interest:

The highest bee and wasp abundances were collected in a plantation

that combined living ground cover along the middle of rows with bare

124

ground along the tree line (Chapter 3). Given our knowledge of the

nesting habits of many hymenopteran species, particularly in

Mediterranean ecosystems, this result is not surprising. A contiguous

carpet of ground cover across the plantation floor can provide diverse

floral resources, but not areas of bare soil for ground-nesting bees and

wasps to excavate their nests into. Hence, almond plantations may be more

capable of supporting permanent pollinator communities if they are

managed with a combination of weedy ground cover and bare soil, rather

than only one of these elements. This provides valuable baseline

information for future research, as previous studies examining

relationships between pollinator taxa and ground cover within fruit

orchards have rarely provided any detail on the level of contiguity of

ground cover vegetation (e.g. Mandelik & Roll 2009), or have only

compared ‘bare’ orchards with ‘weedy’ orchards (e.g. Brown 2012). More

studies that compare pollinator communities across orchards with multiple

different types of ground cover configurations and specifically examine

relationships between these two variables will increase our understanding

of the optimal configuration of this important habitat attribute.

In broadacre monoculture plantations, native bees and hoverflies

were only present in patches of remnant mallee vegetation within and

adjacent to blocks of almond trees (Chapter 4). Native bees and

hoverflies are the most recognised pollinators and this result has

significant potential to motivate changes toward ecological management

practices in the almond industry. It is possible that the sampling method

underrepresented the abundance and richness of the pollinator taxa

collected in plantations, as plentiful floral resources may divert insects

from pan traps (Baum & Wallen 2011) and the single sampling event in

this study could not account for temporal variation in pollinator

abundance. However, the clear differences in abundance and richness of

pollinator taxa caught in the plant-diverse orchards during flowering,

compared to the abundance and richness of the same taxa caught in

monoculture plantations (Chapters 3 & 4), shows that the method was able

to provide comparative information on how pollinators were using the two

plantation types during the critical flowering period. The pattern seen here

125

agrees with other studies suggesting that bees and hoverflies may be

highly-sensitive to severe homogeneity and changes in vegetation contrast

(Wratten et al. 2003; Kohler et al. 2008), as well as Brosi et al.’s (2008)

recommendation to maintain small patches of natural habitat interspersed

through intensively-managed landscapes. A long-term comparison of

pollinator communities across plantations containing different sizes and

configurations of internal mallee patches would shed more light on the

importance of these landscape elements for supporting wild pollinators in

broadacre plantation landscapes.

Of the two monoculture plantations, Wemen plantation had

consistently more abundant and diverse pollinator communities than

Liparoo plantation (Chapters 3-6). Wemen plantation was smaller than

Liparoo, and also had more average ground cover vegetation across the

orchard floor (Chapter 3). The implications of these results have been

discussed in relevant chapters; however, it is likely that the smaller area

gave foraging pollinator insect fauna greater access to the interior of the

plantation and the presence of sparse ground cover also provided a limited

amount of non-crop resources for pollinators. Given the overall similarity

between the two plantations compared to other sampled vegetation types,

this is unlikely to have had a significant impact on my results as a whole,

but the ecological and management implications of this pattern requires

more robust investigation.

In combination, the results presented here provide incentive for further detailed

research programs, because of the study system’s two-fold significance. The

mallee bioregions are unique ecosystems with high conservation value (Mallee

Catchment Management Authority 2008; Cheal et al. 2013), but our knowledge of

flying invertebrate fauna is limited (Yen et al. 2006). Additionally, almond is an

economically-significant crop for Australia’s economy and the current pressures

on the commercial honeybee industry warrant serious investigation of alternative

pollination services available to growers.

Based on the findings presented in this work, the limitations of my

research can be expanded on through a number of detailed studies. In particular,

my sampling method was chosen to maximise sampling time per site and to

minimise bias from inexperience in visual identification and netting techniques

126

(Westphal et al. 2008; Popic et al. 2013). However, my sampling method did not

account for potential Lepidoptera (moths and butterflies) or Formicidae (ant)

pollinators, although individuals of both these groups were seen visiting flowers

during sampling (see Appendix 2) and very small numbers of other known

pollinator insect groups, Thysanoptera (thrips) and Coleoptera (beetles), were also

caught in pan traps during sampling. Furthermore, family-level identification of

insect samples is useful for providing broad estimates of ecosystem service

providers, especially when reliable taxonomic information is limited, as was the

case with my system; however, higher-level taxonomic identification may

increase the application of these results after specific taxa of conservation and

management interest to the system have been identified. Hence, experienced net

sampling of the flowering almond canopy, combined with detailed flower visitor

observations and species-level identification of samples, would expand my results

with a more robust estimate of wild pollinator taxa available to pollinate almond

flowers in this system. In addition to the sampling limitations, the logistical

restrictions on the number of sites I could sample on each trip, particularly in the

mallee vegetation, would have limited the breadth of the pollinator community I

sampled. The accumulation curves shown in Chapter 3 (page 38) indicated that I

had most likely not captured the full representation of wild hymenopteran taxa,

although the curves for mallee dipteran communities and orchard hymenopteran

communities did approach an asymptote. In addition, a number of taxa across all

pollinator groups, and in both years, were only collected at mallee sites (Appendix

2) and the wetter than average weather conditions across the region during the two

sampling years most likely influenced the abundance and richness of my target

insects (Chapter 2). Given the lack of ecological information available on

Hymenoptera and Diptera of the Victorian mallee regions, these are exciting areas

for further investigation.

7.2 Conclusions

The results presented in this thesis suggest that plant-diverse almond plantations

established in heterogeneous landscapes contribute to local biodiversity

conservation, and thus benefit from wild pollinators, more than intensively-

managed plantations cultivated as broadacre monocultures. My results concur

with numerous studies that have shown that: (a) weedy ground cover in crop

127

systems increases the abundance and richness of pollinator taxa (e.g. Mandelik &

Roll 2009; Brown 2012); (b) ecological management practices on individual

farms are often more responsible for creating resource heterogeneity and

connectivity than the presence of particular vegetation types in the surrounding

landscape (e.g. Kleijn & van Langevelde 2006; Persson & Smith 2013); (c)

physical habitat boundaries in agroecosystems are not barriers to wild pollinator

movement, but limited structure, complexity and resources available beyond the

boundary may hinder their movement into homogeneous patches (e.g. Kohler et

al.2008); and (d) the availability of nesting and overwintering resources is crucial

for supporting beneficial insect communities in agroecosystems (e.g. Pfiffner &

Luka 2000) and this is especially pertinent to winter-flowering crops like almond.

Thus, the ability of a plantation to benefit from unmanaged pollination services

during the flowering period may depend on its capacity to support pollinators

throughout the plantation before flowering, i.e. overwintering or larval stages

from the previous season, or after flowering, i.e. during the post-emergence

reproductive stage.

The work presented in this thesis has highlighted ecological differences,

and similarities, between commercial almond plantations and local native mallee

woodlands, with particular relevance to the conservation of wild insect

pollinators. Overall, I have shown that ecologically-managed almond plantations

have the potential to support a diverse wild pollinator community, which has

important implications for both almond production and biodiversity conservation

in the mallee regions. In addition, my research has highlighted a number of gaps

in our knowledge of the ecological role of tree crop plantations and how they

influence pollinator communities. This is an essential area for future research,

given the global expansion of economically-important tree crop plantations and

their potential to contribute to the conservation of multiple ecosystem services in

agricultural landscapes. Most importantly, I have contributed to knowledge of the

pollinator insect fauna of the Victorian mallee, a unique but understudied

ecosystem, and presented data showing how these taxa interact with

agroecosystems in the region. These results will hopefully motivate further

scientific and public discourse and more detailed research into the importance of

conserving wild pollinators, particularly in and around agroecosystems

established in regions of high conservation value.

128

129

PART IV: REFERENCE MATERIAL

130

Literature Cited

Abate, T. (1991) Intercropping and weeding – effects on some natural enemies of African bollworm, Heliothis armigera (Hbn.) (Lep., Noctuidae) in bean fields. Journal of Applied Entomology 112:38-42.

Abrahamczyk, S., Steudel, B., Kessler, M. (2010) Sampling Hymenoptera along a precipitation gradient in tropical forests: the effectiveness of different coloured pan traps. Entomologia Experimentalis et Applicata 137:262-268.

Abrol, D.P. (1988) Pollination studies in almond (Prunus amygdalis L.). Current Science57:397-398.

Adams, P.B., Lawson, S.D. (1993) Pollination in Australian orchids: a critical assessment of the literature 1882-1992. Australian Journal of Botany 41:553-575.

Adler, P.B., Lauenroth, W.K. (2003) The power of time: spatiotemporal scaling of species diversity. Ecology Letters 6:749-756.

Aguiar, M.R., Sala, O.E. (1999) Patch structure, dynamics and implications for the functioning of arid ecosystems. Trends in Ecology and Evolution 14:273-277.

Alcock, J. (1981) Notes on the reproductive behaviour of some Australian Thynnine wasps (Hymenoptera: Tiphiidae). Journal of the Kansas Entomological Society 54:681-693.

Almond Board of Australia (2013) Almond Insights 2012-13. Almond Board of Australia Inc., Berri SA. URL: http://australianalmonds.com.au/industry/resources_stats_reports. Accessed 26 July 2013.

Altieri, M.A. (1999) The ecological role of biodiversity in agroecosystems. Agriculture, Ecosystems and Environment 74:19-31.

Altieri, M.A. (2004) Linking ecologists and traditional farmers in the search for sustainable agriculture. Frontiers in Ecology and Environment 2:35-42.

Altieri, M.A., Letourneau, D.K., Davis, J.R. (1983) Developing sustainable agroecosystems. BioScience 33:45-49.

Altieri, M.A., Nicholls, C.I. (2002) The simplification of traditional vineyard based agroforests in northwestern Portugal: some ecological implications. Agroforestry Systems 56:185-191.

Andersen, A.N. (1983) Species diversity and temporal distribution of ants in the semi-arid mallee region of north-western Victoria. Australian Journal of Ecology8:127-137.

131

Andersen, A.N., Yen, A.L. (1985) Immediate effects of fire on ants in the semi-arid mallee region of northwestern Victoria. Australian Journal of Ecology 10:25-30.

Andersen, A.N., Yen, A.L. (1992) Canopy ant communities in the semi-arid Mallee region of north-western Victoria. Australian Journal of Zoology 40:205-214.

Anderson, M.J. (2001) A new method for non-parametric multivariate analysis of variance. Austral Ecology 26:32-46.

Andrew, N., Rodgerson, L., York, A. (2000) Frequent fuel-reduction burning: the role of logs and associated leaf litter in the conservation of ant biodiversity. Austral Ecology 25:99-107.

Aragona, F.B., Orr, B. (2011) Agricultural intensification, monocultures, and economic failure: the case of onion production in the Tipajara watershed on the eastern slope of the Bolivian Andes. Journal of Sustainable Agriculture 35:467-492.

Armstrong, J.A. (1979) Biotic pollination mechanisms in the Australian Flora – a review. New Zealand Journal of Botany 17:467-508.

Arnan, X., Bosch, J., Comas, L., Gracia, M., Retana, J. (2011) Habitat determinants of abundance, structure and composition of flying Hymenoptera communities in mountain old-growth forests. Insect Conservation and Diversity 4:200-211.

Arthur, A.D., Li, J., Henry, S., Cunningham, S.A. (2010) Influence of woody vegetation on pollinator densities in oilseed Brassica fields in an Australian temperate landscape. Basic and Applied Ecology 11:406-414.

Ashman, T-L., Knight, TM.., Steets, J.A., Amarasekare, P., Burd, M., Campbell, D.R., Dudash, M.R., Johnston, M.O., Mazer, S.J., Mitchell, R.J., Morgan, M.T., Wilson, W.G. (2004) Pollen limitation of plant reproduction: ecological and evolutionary causes and consequences. Ecology 85:2408-2421.

Austin, A.D., Yeates, D.K., Cassis, G., Fletcher, M.J., La Salle, J., Lawrence, J.F., McQuillan, P.B., Mound, L.A., Bickel, D.J., Gullan, P.J., Hales, D.F., Taylor, G.S. (2004) Insects 'Down Under' – Diversity, endemism and evolution of the Australian insect fauna: examples from select orders. Australian Journal of Entomology 43:216-234.

Austin, A.D., Johnson, N.F., Dowton, M. (2005) Systematics, evolution, and biology of scelionid and platygastrid wasps. Annual Review of Entomology 50:553-582.

Australian Nut Industry Council (n.d.) Industry Overview, ANIC. Australian Nut Industry Council, Elanora QLD. URL: http://nutindustry.org.au/ANIC/Industry-Overview.asp. Accessed 15 March 2011.

Azevedo, P.C.B., Lopes, G.N., Fonteles, R.S., Vaconcelos, G.C., Moraes, J.L.P., Rebêlo, J.M.M. (2011) The effect of fragmentation on Phlebotomine communities (Diptera: Psychodidae) in areas of ombrophilous forest in São Luís, State of Maranhão, Brazil. Neotropical Entomology 40:271-277.

132

Bale, J.S., Hayward, S.A.L. (2010) Insect overwintering in a changing climate. The Journal of Experimental Biology 213:980-994.

Barbour, R.C., Otahal, Y., Vaillancourt, R.E., Potts, B.M. (2008) Assessing the risk of pollen-mediated gene flow from exotic Eucalyptus globulus plantations into native eucalypt populations of Australia. Biological Conservation 141:896-907.

Barlow, J., Gardner, T.A., Araujo, I.S., Ávila-Pires, T.C., Bonaldo, A.B., Costa, J.E., Esposito, M.C., Ferreira, L.V., Hawes, J., Hernandez, M.I.M., Hoogmoed, M.S., Leite, R.N., Lo-Man-Hung, N.F., Malcolm, J.R., Martins, M.B., Mestre, L.A.M., Miranda-Santos, R., Nunes-Gutjahr, A.L., Overal, W.L., Parry, L., Peters, S.L., Ribeiro-Junior, M.A., da Silva, M.N.F., da Silva Motta, C., Peres, C.A. (2007) Quantifying the biodiversity value of tropical primary, secondary, and plantation forests. Proceedings of the National Academy of Sciences USA 104:18555-18560.

Bartomeus, I., Winfree, R. (2011) The Circe Principle: are pollinators waylaid by attractive habitats? Current Biology 21:652-654.

Batista Matos, M.C., Sousa-Souto, L., Almeida, R.S., Teodoro, A.V. (2013) Contrasting patterns of species richness and composition of solitary wasps and bees (Insecta: Hymenoptera) according to land-use. Biotropica 45:73-79.

Batley, M., Hogendoorn, K. (2009) Diversity and conservation status of native Australian bees. Apidologie 40:347-354.

Baum, K.A., Wallen, K.E. (2011) Potential bias in pan trapping as a function of floral abundance. Journal of the Kansas Entomological Society 84:155-159.

Bejan, A., Lorente, S. (2010) The constructal law of design and evolution in nature. Philosophical Transactions of the Royal Society B 365:1335-1347.

Bennett, A.F., Radford, J.Q., Haslem, A. (2006) Properties of land mosaics: implications for nature conservation in agricultural environments. Biological Conservation133:250-264.

Bennett, E.M., Peterson, G.D., Gordon, L.J. (2009) Understanding relationships among multiple ecosystem services. Ecology Letters 12:1394-1404.

Bennett, A.J., Bending, G.D., Chandler, D., Hilton, S. & Mills, P. (2012) Meeting the demand for crop production: the challenge of yield decline in crops grown in short rotations. Biological Reviews 87:52-71.

Benton, T.G., Vickery, J.A., Wilson, J.D. (2003) Farmland biodiversity: is habitat heterogeneity the key? Trends in Ecology and Evolution 18:182-188.

Bianchi, F.J.J.A., Schellhorn, N.A., Buckley, Y.M., Possingham, H.P. (2010) Spatial variability in ecosystem services: simple rules for predator-mediated pest suppression. Ecological Applications 20:2322-2333.

133

Biesmeijer, J.C., Roberts, S.P.M., Reemer, M., Ohlemüller, R., Edwards, M., Peeters, T., Schaffers, A.P., Potts, S.G., Kleukers, R., Thomas, C.D., Settele, J., Kunin, W.E. (2006) Parallel declines in pollinators and insect-pollinated plants in Britain and the Netherlands. Science 313:351-354.

Blakers, M., Macmillan, L. (1988) Mallee Conservation in Victoria. Research Paper No. 6. Faculty of Environmental Design and Construction, Royal Melbourne Institute of Technology, Melbourne.

Blanche, K.R., Ludwig, J.A., Cunningham, S.A. (2006) Proximity to rainforest enhances pollination and fruit set in orchards. Journal of Applied Ecology 43:1182-1187.

Blitzer, E.J., Dormann, C.F., Holzschuh, A., Klein, A-M., Rand, T.A., Tscharntke, T. (2008) Spillover of functionally important organisms between managed and natural habitats. Agriculture, Ecosystems and Environment 146:34-43.

Bommarco, R., Kleijn, D. & Potts, S.G. (2013) Ecological intensification: harnessing ecosystem services for food security. Trends in Ecology and Evolution 28:230-238.

Bond, W.J. (1994) Do mutualisms matter? Assessing the impact of pollinator and disperser disruption on plant extinction. Philosophical Transactions of the Royal Society B 344:83-90.

Bos, M.M., Veddeler, D., Bogdanski, A.K., Klein, A-M., Tscharntke, T., Steffan-Dewenter, I., Tylianakis, J.M. (2007) Caveats to quantifying ecosystem services: fruit abortion blurs benefits from crop pollination. Ecological Applications17:1841-1849.

Bosch, J., Blas, M. (1994) Foraging behaviour and pollinating efficiency of Osmia cornuta and Apis mellifera on almond (Hymenoptera, Megachilidae and Apidae). Applied Entomology and Zoology 29:1-9.

Bosch, J., Retana, J., Cerdá, X. (1997) Flowering phenology, floral traits and pollinator composition in a herbaceous Mediterranean plant community. Oecologia109:583-591.

Bosch, J., Kemp, W.P. (2002) Developing and establishing bee species as crop pollinators: the example of Osmia spp. (Hymenoptera: Megachilidae) and fruit trees. Bulletin of Entomological Research 92:3-16.

Brittain, C., Kremen, C., Klein, A-M. (2013) Biodiversity buffers pollination from changes in environmental conditions. Global Change Biology 19:540-547.

Brosi, B.J., Armsworth, P.R., Daily, G.C. (2008) Optimal design of agricultural landscapes for pollination services. Conservation Letters 1:27-36.

Brown, M.W. (1999) Applying principles of community ecology to pest management in orchards. Agriculture, Ecosystems and Environment 73:103-106.

134

Brown, M.W. (2012) Role of biodiversity in intergrated fruit production in eastern North American orchards. Agricultural and Forest Entomology 14:89-99.

Brown, A.O., McNeil, J.N. (2009) Pollination ecology of the high-latitude, dioecious cloudberry (Rubus chamaemorus; Rosaceae). American Journal of Botany96:1096-1107.

Bugg, R.L., Waddington, C. (1994) Using cover crops to manage arthropod pests of orchards: a review. Agriculture, Ecosystems and Environment 50:11-28.

Bureau of Meteorology (2013) Climate Data Online (Annuello Station 76000). URL: http://www.bom.gov.au/jsp/ncc/cdio/weatherData/av?p_nccObsCode=139&p_display_type=dataFile&p_startYear=&p_c=&p_stn_num=076000. Accessed January 26 2013.

Burel, F., Baudry, J., Butet, A., Clergeau, P., Delettre Y., Le Coeur D., Dubs F., Morvan N., Paillat G., Petit S., Thenail C., Brunel E., Lefeuvre J-C. (1998) Comparative biodiversity along a gradient of agricultural landscapes. Acta Oecologia 19:47-60.

Burnham, K.P., Anderson, D.R. (2002) Model Selection and Multimodel Inference. Springer Science, New York.

Buzas, M.A., Hayek, L-A.C. (1996) Biodiversity resolution: An integrated approach. Biodiversity Letters 3:40-43.

Buzas, M.A., Hayek, L-A.C. (1998) SHE analysis for biofacies identification. Journal of Foraminiferal Research 28:233-239.

Buzas, M.A., Hayek, L-A.C. (2005) On richness and evenness within and between communities. Paleobiology 31:199-220.

Cadenasso, M.L., Pickett, S.T.A. (2000) Linking forest edge structure to edge function: mediation of herbivore damage. Journal of Ecology 88:31-44.

Cadenasso, M.L., Pickett, S.T.A., Weathers, K.C., Bell, S.S., Benning, T.L., Carreiro, M.M., Dawson, T.E. (2003) An interdisciplinary and synthetic approach to ecological boundaries. BioScience 53:717-722.

Cadotte, M.W., Carscadden, K., Mirotchnick, N. (2011) Beyond species: functional diversity and the maintenance of ecological processes and services. Journal of Applied Ecology 48:1079-1087.

Calderone, N.W. (2012) Insect pollinated crops, insect pollinators and US agriculture: trend analysis of aggregate data for the period 1992-2009. PLoS ONE7(5):e37235. doi:10.1371/journal.pone.0037235.

Cale, P.G., Hobbs, R.J. (1994) Landscape heterogeneity indices: problems of scale and applicability, with particular reference to animal habitat description. Pacific Conservation Biology 1:183-93.

135

Campbell, R.E., Harding, J.S., Ewers, R.M., Thorpe, S., Didham, R.K. (2011) Production land use alters edge response functions in remnant forest invertebrate communities. Ecological Applications 21:3147-3161.

Cane, J.H., Minckley, R.L., Kervin, L.J. (2000) Sampling bees (Hymenoptera: Apiformes) for pollinator community studies: pitfalls of pan trapping. Journal of the Kansas Entomological Society 73:225-231.

Cardoso, P., Erwin, T.L., Borges, P.A.V., New, T.R. (2011) The seven impediments in invertebrate conservation and how to overcome them. Biological Conservation144:2647-2655.

Carreck, N.L., Williams, I.H. (2002) Food for insect pollinators on farmland: insect visits to flowers of annual seed mixtures. Journal of Insect Conservation 6:13-23.

Carvalheiro, L.G., Seymour, C.L., Veldtman, R., Nicolson, S.W. (2010) Pollination services decline with distance from natural habitat even in biodiversity-rich areas. Journal of Applied Ecology 47:810-820.

Carvalheiro, L.G., Veldtman, R., Shenkute, A.G., Tesfay, G.B., Walter, C., Pirk, W., Donaldson, J.S., Nicolson, S.W. (2011) Natural and within-farmland biodiversity enhances crop productivity. Ecology Letters 14:251-259.

Chacoff, N.P., Aizen, M.A. (2006) Edge effects on flower-visiting insects in grapefruit plantations bordering premontane subtropical forest. Journal of Applied Ecology43:18-27.

Chapman, M.G., Underwood, A.J. (1999) Ecological patterns in multivariate assemblages: information and interpretation of negative values in ANOSIM tests. Marine Ecology Progress Series 180:257-265.

Cheal, D.C., Parkes, D.M. (1989) Mallee Vegetation in Victoria. In: J.C. Noble, R.A. Bradstock (eds.) Mediterranean Landscapes in Australia: Mallee Ecosystems and their Management, pp. 125-140. CSIRO, East Melbourne.

Cheal, D., Moxham, C., Kenny, S., Millet-Riley, J. (2013) Rare plant recovery in Mallee woodlands. The Victorian Naturalist 130:96-108.

Clarke, K.R. (1993) Non-parametric multivariate analyses of changes in community structure. Australian Journal of Ecology 18:117-143.

Clarke, K.R., Ainsworth, M. (1993) A method of linking multivariate community structure to environmental variables. Marine Ecology Progress Series 92:205-219.

Cole, P.J., Sharley, A., Wetherby, K.G. (1989) Irrigated Agriculture. In: J.C. Noble, R.A. Bradstock (eds.) Mediterranean Landscapes in Australia: Mallee Ecosystems and their Management, pp. 329-342. CSIRO, East Melbourne.

136

Collinge, S.K., Palmer, T.M. (2002) The influences of patch shape and boundary contrast on insect response to fragmentation in California grasslands. Landscape Ecology17:647-656.

Collinge, S.K., Prudic, K.L., Oliver, J.C. (2006) Effects of local habitat characteristics and landscape context on grassland butterfly diversity. Conservation Biology17:178-187.

Colloff, M.J., Fokstuen, G., Boland, T. (2003) Toward the Triple Bottom Line in sustainable horticulture: biodiversity, ecosystem services and an environmental management system for citrus orchards in the Riverland of South Australia.CSIRO Entomology, Canberra, ACT.

Colwell, R.K. (2013) EstimateS: Statistical estimation of species richness and shared species from samples. Version 9.1. User's Guide and application published at: http://purl.oclc.org/estimates. Accessed 30 June 2013.

Colwell, R.K., Coddington, J.A. (1994) Estimating biodiversity through extrapolation. Philosophical Transactions of the Royal Society B 345:101-118.

Colwell, R.K., Mao, C.X., Chang, J. (2004) Interpolating, extrapolating, and comparing incidence-based species accumulation curves. Ecology 85:2717-2727.

Colwell, R.K., Chao, A., Gotelli, N.J., Lin, S-Y., Mao, C.X., Chazdon, R.L., Longino, J.T. (2012) Models and estimators linking individual-based and sample-based rarefactio, extrapolation and comparison of assemblages. Journal of Plant Ecology 5:3-21.

Concepción, E.D., Díaz, M., Kleijn, D., Báldi, A., Batáry, P., Clough, Y., Gabriel, D., Herzog, F., Holzschuh, A., Knop, E., Marshall, E.J.P., Tscharntke, T., Verhulst, J. (2012) Interactive effects of landscape context constrain the effectiveness of local agri-environmental management. Journal of Applied Ecology 49:695-705.

Coudrain, V., Herzog, F., Entling, M.H. (2013) Effects of habitat fragmentation on abundance, larval food and parasitism of a spider-hunting wasp. PLOS One8(3):e59286. doi:10.1371/journal.pone.0059286.

Cunningham, S.A. (2000) Effects of habitat fragmentation on the reproductive ecology of four plant species in mallee woodland. Conservation Biology 14:758-768.

Cunningham, S.A., FitzGibbon, F., Heard, T.A. (2002) The future of pollinators for Australian agriculture. Australian Journal of Agricultural Research 53:893-900.

Dąbrowska-Prot, E., Wasiłowska, A. (2012) The role of ecotones in man-disturbed landscape: boundaries between mixed forest and adjacent man-made ecosystems in the Kampinos National Park, Poland. Polish Journal of Ecology 60:677-698.

Dafni, A. (1996) Autumnal and winter pollination adaptations under Mediterranean conditions. Bocconea 5:171-181.

137

Daily, G.C. (1997) Nature’s Services: Societal dependence on natural ecosystems. Island Press, Washington D.C.

Dangerfield, J.M., Pik, A.J., Britton, D., Holmes, A., Gillings, M., Oliver, I., Briscoe, D., Beattie, A. (2003) Patterns of invertebrate biodiversity across a natural edge. Austral Ecology 28:227-236.

Degrandi-Hoffman, G. (2001) The pollination of almonds. American Bee Journal141:655-657.

Delettre, Y.R. (2005) Short-range spatial patterning of terrestrial Chironomidae (Insecta: Diptera) and farmland heterogeneity. Pedobiologia 49:15-27.

Department of Sustainability and Environment (2013) Conservation and Environment: Victoria’s Bioregions. The State of Victoria, Australia. URL: http://www.dse.vic.gov.au/conservation-and-environment/native-vegetation-groups-for-victoria/victorias-bioregions. Accessed 4 September 2013.

Diacon-Bolli, J., Dalang, T., Holderegger, R., Bürgi, M. (2012) Heterogeneity fosters biodiversity: linking history and ecology of dry calcareous grasslands. Basic and Applied Ecology 13:641-653.

D’Odorico, P., Laio, F., Porporato, A., Ridolfi, L., Rinaldo, A., Rodriguez-Iturbe, I. (2010) Ecohydrology of terrestrial ecosystems. BioScience 60:898-907.

Driscoll, D.A., Weir, T. (2005) Beetle responses to habitat fragmentation depend on ecological traits, habitat condition, and remnant size. Conservation Biology19:182-194.

Dubois, G.F., Vignon, V., Delettre, Y.R., Rantier, Y., Vernon, P., Burel, F. (2009) Factors affecting the occurrence of the endangered saproxylic beetle Osmoderma eremita (Scopoli, 1763) (Coleoptera: Cetoniidae) in an agricultural landscape. Landscape and Urban Planning 91:152-159.

Duncan, D.H., Dorrough, J.W. (2009) Historical and current land use shape landscape restoration options in the Australian wheat and sheep farming zone. Landscape and Urban Planning 91:124-132.

Dunning, J.B., Danielson, B.J., Pulliam, H.R. (1992) Ecological processes that affect populations in complex landscapes. Oikos 65:169-175.

Dutoit, T., Buisson, E., Gerbaud, E., Roche, P., Tatoni, T. (2007) The status of transitions between cultivated fields and their boundaries: ecotones, ecoclines or edge effects? Acta Oecologia 31:127-136.

Dyer, A.G., Boyd-Gerny, S., McLoughlin, S., Rosa, M.G.P., Simonov, V., Wong, B.B.M. (2012) Parallel evolution of angiosperm colour signals: common evolutionary pressures linked to hymenopteran vision. Proceedings of the Royal Society B279:3606-3615.

138

Ebeling, A., Klein, A-M., Schumacher, J., Weisser, W.W., Tscharntke, T. (2008) How does plant richness affect pollinator richness and temporal stability of flower visits? Oikos 117:1808-1815.

Ebeling, A., Klein, A-M., Tscharntke, T. (2011) Plant-flower interaction webs: Temporal stability and pollinator specialization increases along an experimental plant diversity gradient. Basic and Applied Ecology 12:300-309.

Ehrlich, P.R., Mooney, H.A. (1983) Extinction, substitution and ecosystem services. BioScience 33:248-254.

Eilers, E.J., Klein, A-M. (2009) Landscape context and management effects on an important insect pest and its natural enemies in almond. Biological Control51:388-394.

Ekroos, J., Heliölä, J., Kuussaari, M. (2010) Homogenisation of lepidopteran communities in intensively cultivated agricultural landscapes. Journal of Applied Ecology 47:459-467.

Elberling, H., Olesen, J.M. (1999) The structure of a high-latitude plant-flower visitor system: the dominance of flies. Ecography 22:314-323.

Essenberg, C.J. (2013) Scale-dependent shifts in the species composition of flower visitors with changing floral density. Oecologia 171:187-196.

Ewers, R.M., Didham, R.K. (2006) Continuous response functions for quantifying the strength of edge effects. Journal of Applied Ecology 42:527-536.

Fægri, K., van der Pijl, L. (1979) The Principles of Pollination Ecology. Pergamon Press Ltd., Oxford.

Fagan, W.F., Cantrell, R.S., Cosner, C. (1999) How habitat edges change species interactions. The American Naturalist 153:165-182.

Faith, D.P., Minchin, P.R., Belbin, L. (1987) Compositional dissimilarity as a robust measure of ecological distance. Vegetatio 69:57-68.

Farina, A. (1998) Principles and Methods in Landscape Ecology. Chapman & Hall, London.

Fayt, P., Dufrêne, M., Branquart, E., Hastir, P., Pontégnie, C., Henin, J-M., Versteirt, V. (2006) Contrasting responses of saproxylic insects to focal habitat resources: the example of longhorn beetles and hoverflies in Belgian deciduous forests. Journal of Insect Conservation 10:129-150.

Fletcher Jr., R.J., Ries, L., Battin, J., Chalfoun, A.D. (2007) The role of habitat area and edge in fragmented landscapes: definitively distinct or inevitably intertwined? Canadian Journal of Zoology 85:1017-1030.

139

Fonseca, C.R., Joner, F. (2007) Two-sided edge effect studies and the restoration of endangered ecosystems. Restoration Ecology 15:613-619.

Ford, H.A., Forde, N. (1976) Birds as possible pollinators of Acacia pycnantha. Australian Journal of Botany 24:793-795.

Forister, M.L. (2009) Anthropogenic islands in the arid west: comparing the richness and diversity of insect communities in cultivated fields and neighboring wildlands.Environmental Entomology 38:1028-1037.

Forman, R.T.T., Godron, M. (1986) Landscape Ecology. John Wiley & Sons, New York.

Francis, C.A., Porter, P. (2011) Ecology in sustainable agriculture practices and systems. Critical Reviews in Plant Sciences 30:64-73.

Freeman Long, R., Corbett, A., Lamb, C., Reberg-Horton, C., Chandler, J., Stimmann, M. (1998) Beneficial insects move from flowering plants to nearby crops. California Agrriculture 52:23-26.

Friend, G.R., Williams, M.R. (1996) Impact of fire on invertebrate communities in mallee-heath shrublands of southwestern Australia. Pacific Conservation Biology2:244-267.

Fründ, J., Linsenmair, K.E., Blüthgen, N. (2010) Pollinator diversity and specialization in relation to flower diversity. Oikos 119:1581-1590.

Gaffney, A., Allen, G.R., Brown, P.H. (2011) Insect visitation to flowering hybrid carrot seed crops. New Zealand Journal of Crop and Horticultural Science 39:79-93.

Gámez-Virués, S., Gurr, G., Raman, A., La Salle, J., Nicol, H. (2009) Effects of flowering groundcover vegetation on diversity and activity of wasps in a farm shelterbelt in temperate Australia. BioControl 54:211-218.

Gámez-Virués, S., Gurr, G.M., Raman, A., Nicol, H.I. (2010) Plant diversity and habitat structure affect tree growth, herbivory and natural enemies in shelterbelts. Basic and Applied Ecology 11:542-549.

Garibaldi, L.A., Steffan-Dewenter, I., Kremen, C., Morales, J.M., Bommarco, R., Cunningham, S.A., Carvalheiro, L.G., Chacoff, N.P., Dudenhöffer, J.H., Greenleaf, S.S., Holzschuh, A., Isaacs, R., Krewenka, K., Mandelik, Y., Mayfield, M.M., Morandin, L.A., Potts, S.G., Ricketts, T.H., Szentgyörgyi, H., Viana, B.F., Westphal, C., Winfree, R., Klein, A-M. (2011) Stability of pollination services decreases with isolation from natural areas despite honey bee visits. Ecology Letters 14:1062-1072.

Garibaldi, L.A., Steffan-Dewenter, I., Winfree, R., Aizen, M.A., Bommarco, R., Cunningham, S.A., Kremen, C., Carvalheiro, L.G., Harder, L.D., Afik, O., Bartomeus, I., Benjamin, F., Boreux, V., Cariveau, D., Chacoff, N.P., Dudenhöffer, J.H., Freitas, B.M., Ghazoul, J., Greenleaf, S., Hipólito, J., Holzschuh, A., Howlett, B., Isaacs, R., Javorek, S.K., Kennedy, C.M., Krewenka,

140

K.M., Krishnan, S., Mandelik, Y., Mayfield, M.M., Motzke, I., Munyuli, T., Nault, B.A., Otieno, M., Petersen, J., Pisanty, G., Potts, S.G., Rader, R., Ricketts, T.H., Rundlöf, M., Seymour, C.L., Schüepp, C., Szentgyörgyi, H., Taki, H., Tscharntke, T., Vergara, C.H., Viana, B.F., Wanger, T.C., Westphal, C., Williams, N., Klein, A-M. (2013) Wild pollinators enhance fruit set of crops regardless of honey bee abundance. Science 339:1608-1611.

Gattone, S.A., Di Battista, T. (2009) A functional approach to diversity profiles. Journal of Applied Statistics 58:267-284.

Gemmill-Herren, B., Ochieng, A.O. (2008) Role of native bees and natural habitats in eggplant (Solanum melongena) pollination in Kenya. Agriculture, Ecosystems and Environment 127:31-36.

Gittings, T., O’Halloran, J., Kelly, T., Giller, P.S. (2006) The contribution of open spaces to the maintenance of hoverfly (Diptera, Syrphidae) biodiversity in Irish plantation forests. Forest Ecology and Management 237:290-300.

Gómez, J.A., Llewellyn, C., Basch, G., Sutton, P.B., Dyson, J.S., Jones, C.A. (2011) The effects of cover crops and conventional tillage on soil and runoff loss in vineyards and olive groves in several Mediterranean countries. Soil Use and Management27:502-514.

Gomiero, T., Pimentel, D., Paoletti, M.G. (2011) Environmental impact of different agricultural management practices: conventional vs. organic agriculture. Critical Reviews in Plant Sciences 30:95-124.

Gotelli, N.J., Colwell, R.K. (2001) Quantifying biodiversity: procedures and pitfalls in the measurement and comparison of species richness. Ecology Letters 4:379-391.

Gotelli, N.J., Colwell, R.K. (2011) Estimating species richness. In: A.E. Magurran, B.J. McGill (eds.) Biological Diversity: Frontiers in measurement and assessment, pp. 39-54. Oxford University Press, Oxford.

Grasswitz, T.R., James, D.G. (2009) Influence of hop yard ground flora on invertebrate pests of hops and their natural enemies. Journal of Applied Entomology 133:210-221.

Greenleaf, S.S., Williams, N.M., Winfree, R., Kremen, C. (2007) Bee foraging ranges and their relationship to body size. Oecologia 153:589-596.

Griggs, W.H., Iwakiri, B.T. (1975) Pollen tubes growth in almond flowers. CaliforniaAgriculture 29:4-7.

Gruber, B., Eckel, K., Everaars, J., Dormann, C.F. (2011) On managing the red mason bee (Osmia bicornis) in apple orchards. Apidologie 42:564-576.

Grundel, R., Jean, R.P., Frohnapple, K.J., Glowacki, G.A., Scott, P.E., Pavlovic, N.B. (2010) Floral and nesting resources, habitat structure, and fire influence bee

141

distribution across an open-forest gradient. Ecological Applications 20:1678-1692.

Haaland, C., Naisbit, R.E., Bersier, L-F. (2010) Sown wildflower strips for insect conservation: a review. Insect Conservation and Diversity 4:60-80.

Haenke, S., Scheid, B., Schaefer, M., Tscharntke, T., Thies, T. (2009) Increasing syrphid fly diversity and density in sown flower strips within simple vs. complex landscapes. Journal of Applied Ecology 46:1106-1114.

Hagan, M., Kraemer, M. (2010) Agricultural surroundings support flower-visitor networks in an Afrotropical rain forest. Biological Conservation 143:1654-1663.

Hamilton, J., Yeates, D., Hastings, D., Colless, D., McAlpine, D., Bickel, D., Daniels, G., Schneider, M., Cranston, P., Marshall, S. (2006) On the Fly: the interactive atlas and key to Australian fly families (CD-ROM). Australian Biological Resources Study, Canberra, ACT.

Hammer, Ø., Harper, D.A.T., Ryan, P.D. (2001) PAST: Paleontological statistics software package for education and data analysis. Paeleontologia Electronica4(1):9pp. URL: http://palaeo-electronica.org/2001_1/past/issue1_01.htm. Accessed November 14 2010.

Harding, J.S., Benfield, E.F., Bolstad, P.V., Helfman, G.S., Jones III, E.B.D. (1998) Stream biodiversity: the ghost of land use past. Proceedings of the National Acadamy of Sciences USA 95:14843-14847.

Hartemink, A.E. (2005) Plantation agriculture in the tropics. Outlook on Agriculture34:11-21.

Hartley, M.J. (2002) Rationale and methods for conserving biodiversity in plantation forests. Forest Ecology & Management 155:81-95.

Haslem, A., Bennett, A.F. (2008) Birds in agricultural mosaics: the influence of landscape pattern and countryside heterogeneity. Ecological Applications 18:185-196.

Haslett, J.R. (2001) Biodiversity and conservation of Diptera in heterogeneous land mosaics: a fly's eye view. Journal of Insect Conservation 5:71-75.

Hawkeswood, T.J. (1982) Notes on insect pollination of two species of Eucalyptus(Myrtaceae) from south-west Western Australia. The Victorian Naturalist 99:28-37.

Hayek, L-A.C., Buzas, M.A. (1997) Surveying Natural Populations. Columbia University Press, New York.

Heard, T.A. (1999) The role of stingless bees in crop pollination. Annual Review of Entomology 44:183-206.

142

Heard, T.A., Vithanage, V., Chacko, E.K. (1990) Pollination biology of cashew in the Northern Territory of Australia. Australian Journal of Agricultural Research41:1101-1114.

Heard, T.A., Exley, E.M. (1994) Diversity, abundance, and distribution of insect visitors to macadamia flowers. Environmental Entomology 23:91-100.

Henry, M., Béguin, M., Requier, F., Rollin, O., Odoux, J-F., Aupinel, P., Aptel, J., Tchamitchian, S., Decourtye, A. (2012) A common pesticide decreases foraging success and survival in honey bees. Science 336:348-350.

Herrmann, J.D., Kormann, U., Schüepp, C., Stocker, Y., Herzog, F., Entling, M.H. (2012) Effects of habitat isolation and predation pressure on an arboreal food-web. Community Ecology 13:82-87.

Hill, M.O. (1973) Diversity and evenness: a unifying notation and its consequences. Ecology 54:427-432.

Hill, S.J., Stephenson, D.W., Taylor, B.K. (1985) Almond pollination studies: pollen production and viability, flower emergence and cross-pollination tests. Australian Journal of Experimental Agriculture 25:697-704.

Hoehn, P., Steffan-Dewenter, I., Tscharntke, T. (2010) Relative contribution of agroforestry, rainforest and openland to local and regional bee diversity. Biodiversity and Conservation 19:2189-2200.

Holland, P.G. (1969) The plant patterns of different seasons in two stands of mallee vegetation. Journal of Ecology 57:323-333.

Holland, R.A., Eigenbrod, F., Armsworth, P.R., Anderson, B.J., Thomas, C.D., Gaston, K.J. (2011) The influence of temporal variation on relationships between ecosystem services. Biodiversity and Conservation 20:3285-3294.

Holzschuh, A., Steffan-Dewenter, I., Tscharntke, T. (2008) Agricultural landscapes with organic crops support higher pollinator diversity. Oikos 117:354-361.

Holzschuh, A., Steffan-Dewenter, I., Tscharntke, T. (2010) How do landscape composition and configuration, organic farming and fallow strips affect the diversity of bees, wasps and their parasiotids? Journal of Animal Ecology 79:491-500.

Holzschuh, A., Dormann, C.F., Tscharntke, T., Steffan-Dewenter, I. (2011) Expansion of mass-flowering crops leads to transient pollinator dilution and reduced wild plant pollination. Proceedings of the Royal Society B 278:3444-3451.

Holzschuh, A., Dormann, C.F., Tscharntke, T., Steffan-Dewenter, I. (2013) Mass-flowering crops enhance wild bee abundance. Oecologia 172:477-484.

Hooks, C.R.R., Johnson, M.W. (2004) Using undersown clovers as living mulches: effects on yields, lepidopterous pest infestations, and spider densities in a

143

Hawaiian broccoli agroecosystem. International Journal of Pest Management50:115-120.

Horskins, K., Turner, V.B. (1999) Resource use and foraging patterns of honeybees, Apis mellifera, and native insects on flowers of Eucalyptus corsata. Australian Journal of Ecology 24:221-227.

Horton, D.R., Broers, D.A., Lewis, R.R., Granatstein, D., Zack, R.S., Unruh, T.R., Moldenke, A.R., Brown, J.J. (2003) Effects of mowing frequency on densities of natural enemies in three Pacific Northwest pear orchards. Entomologia Experimentalis et Applicata 106:135-145.

Houston, T.F., Maynard, G.V. (2012) An unusual new paracolletine bee, Leioproctus (Ottocolletes) muelleri subgen. & sp. nov. (Hymenoptera: Colletidae): with notes on nesting biology and in-burrow nest guarding by macrocephalic males. Australian Journal of Entomology 51:248-257.

Humphrey, J.W., Hawes, C., Peace, A.J., Ferris-Kaan, R., Jukes, M.R. (1999) Relationships between insect diversity and habitat characteristics in plantation forests. Forest Ecology and Management 113:11-21.

Hunter, M.D. (2002) Landscape structure, habitat fragmentation, and the ecology of insects. Agricultural and Forest Entomology 4:159-166.

Hyvönen, T., Huusela-Veistola, E. (2008) Arable weeds as indicators of agricultural intensity – a case study from Finland. Biological Conservation 141:2857-2864.

IBM Corporation (2011) IBM SPSS Statistics Version 20. IBM Corporation, Somers, NY.

Jackson, D., Looney, N., Morley-Bunker, M., Thiele, G. (eds.) (2011) Temperate and subtropical fruit production. 3rd edn. CABI International, Oxfordshire UK.

Jackson, J.F. (1996) Gene flow in pollen in commercial almond orchards. Sexual Plant Reproduction 9:367-369.

Jauker, F., Diekötter, T., Schwarzbach, F., Wolters, V. (2009) Pollinator dispersal in an agricultural matrix: opposing responses of wild bees and hoverflies to landscape structure and distance from main habitat. Landscape Ecology 24:547-555.

Jauker, F., Peter, F., Wolters, V., Diekötter, T. (2012) Early reproductive benefits of mass-flowering crops to the solitary bee Osmia rufa outbalance post-flowering disadvantages. Basic and Applied Ecology 13:268-276.

Jha, S., Vandermeer, J.H. (2009) Contrasting bee foraging in response to resource scale and local habitat management. Oikos 118:1174-1180.

Jha, S., Kremen, C. (2013) Resource diversity and landscape-level homogeneity drive native bee foraging. Proceedings of the National Academy of Sciences USA 110:555-558.

144

Jones, G.A., Gillett, J.L. (2005) Intercropping with sunflowers to attract beneficial insects in organic agriculture. Florida Entomologist 88:91-96.

Jordano, P. (1993) Pollination biology of Prunus mahaleb L.: deferred consequences of gender variation for fecundity and seed size. Biological Journal of the Linnean Society 50:65-84.

José-María, L., Armengot, L., Blanco-Moreno, J.M., Bassa, M., Sans, F.X. (2010) Effects of agricultural intensification on plant diversity in Mediterranean dryland cereal fields. Journal of Applied Ecology 47:832-840.

Jost, L. (2006) Entropy and diversity. Oikos 113:363-375.

Jules, E.S., Shahani, P. (2003) A broader ecological context to habitat fragmentation: why matrix habitat is more important than we thought. Journal of Vegetation Science14:459-464.

Kalske, A., Muola, A., Laukkanen, L., Mutikainen, P., Leimu, R. (2012) Variation and constraints of local adaptation of a long-lived plant, its pollinators and specialist herbivores. Journal of Ecology 100:1359-1372.

Kearns, C.A. (2001) North American Dipteran pollinators: assessing their value and conservation status. Conservation Ecology 5:article 5. URL: http://www.consecol.org/vol5/iss1/art5/. Accessed 11 January 2013.

Kehinde, T., Samways, M.J. (2012) Endemic pollinator response to organic vs. conventional farming and landscape context in the Cape Floristic Region biodiversity hotspot. Agriculture, Ecosystems and Environment 146:162-167.

Kennedy, C.M., Lonsdorf, E., Neel, M.C., Williams, N.M., Ricketts, T.H., Winfree, R., Bommarco, R., Brittain, C., Burley, A.L., Cariveau, D., Carvalheiro, L.G., Chacoff, N.P., Cunningham, S.A., Danforth, B.A., Dudenhöffer, J-H., Elle, E., Gaines, H.R., Garibaldi, L.A., Gratton, C., Holzschuh, A., Isaacs, R., Javorek, S.K., Jha, S., Klein, A-M., Krewenka, K., Mandelik, Y., Mayfield, M.M., Morandin, L., Neame, L.A., Otieno, M., Park, M., Potts, S.G., Rundlöf Saez, A., Steffan-Dewenter, I., Taki, H., Viana, B.F., Westphal, C., Wilson, J.K., Greenleaf, S.S., Kremen, C. (2013) A global quantitative synthesis of local and landscape effects on wild bee pollinators in agroecosystems. Ecology Letters 16:584-599.

Kester, D.E., Ross, N.W. (1996) History. In W.C. Micke (Ed.) Almond Production Manual. Division of Agriculture and Natural Resources, University of California, Oakland.

Kevan, P.G. (1973) Parasitoid wasps as flower visitors in the Canadian high Arctic. Journal of Pest Science 46:3-7.

Kevan, P.G. (1999) Pollinators as bioindicators of the state of the environment: species, activity and diversity. Agriculture, Ecosystems and Environment 74:373-393.

145

Kevan, P.G., Baker, H.G. (1983) Insects as flower visitors and pollinators. Annual Review of Entomology 28:407-453.

Kiers, E.T., Palmer, T.M., Ives, A.R., Bruno, J.F., Bronstein, J.L. (2010) Mutualisms in a changing world: an evolutionary perspective. Ecology Letters 13:1459-1474.

King, L.E., Douglas-Hamilton, I., Vollrath, F. (2011) Beehive fences as effective deterrents for crop-raiding elephants: field trials in northern Kenya. African Journal of Ecology 49:431-439.

Kleijn, D., van Langevelde, F. (2006) Interacting effects of landscape context and habitat quality on flower visiting insects in agricultural landscapes. Basic and Applied Ecology 7:201-214.

Klein, A-M., Steffan-Dewenter, I., Tscharntke, T. (2004) Foraging trip duration and density of megachilid bees, eumenid wasps and pompilid wasps in tropical agroforestry systems. Journal of Animal Ecology 73:517-525.

Klein, A-M., Vaissière, B.E., Cane, J.H., Steffan-Dewenter, I., Cunningham, S.A., Kremen, C., Tscharntke, T. (2007) Importance of pollinators in changing landscapes for world crops. Proceedings of the Royal Society B 274:303-313.

Klein, A-M., Brittain, C., Hendrix, S.D., Thorp, R., Williams, N., Kremen, C. (2012) Wild pollination services to California almond rely on semi-natural habitat. Journal of Applied Ecology 49:723-732.

Knerer, G., Schwarz, M. (1976) Halictine social evolution: the Australian enigma. Science 194:445-448.

Knight, T.M., Steets, J.A., Vamosi, J.C., Mazer, S.J., Burd, M., Campbell, D.R., Dudash, M.R., Johnston, M.O., Mitchell, R.J., Ashman, T-L. (2005) Pollen limitation of plant reproduction: pattern and process. Annual Review of Ecology, Evolution & Systematics 36:467-97.

Kohler, F., Verhulst, J., van Klink, R., Kleijn, D. (2008) At what spatial scale do high-quality habitats enhance the diversity of forbs and pollinators in intensively farmed landscapes? Journal of Applied Ecology 45:753-762.

Kovács-Hostyánszki, A., Batáry, P., Báldi, A. (2011) Local and landscape effects on bee communities of Hungarian winter cereal fields. Agricultural and Forest Entomology 13:59-66.

Krebs, J.R., Wilson, J.D., Bradbury, R.B., Siriwardena, G.M. (1999) The second Silent Spring? Nature 400:611-612.

Krell, F-T. (2004) Parataxonomy vs. taxonomy in biodiversity studies – pitfalls and applicability of ‘morphospecies’ sorting. Biodiversity and Conservation 13:795-812.

146

Kremen, C. (2005) Managing ecosystem services: what do we need to know about their ecology? Ecology Letters 8:468-479.

Kremen, C., Williams, N.M., Aizen, M.A., Gemmill-Herren, B., LeBuhn, G., Minckley, R., Packer, L., Potts, S.G., Roulston, T., Steffan-Dewenter, I., Vázquez, D.P., Winfree, R., Adams, L., Crone, E.E., Greenleaf, S.S., Keitt, T.H., Klein, A-M., Regetz, J., Ricketts, T.H. (2007) Pollination and other ecosystem services provided by mobile organisms: a conceptual framework for the effects of land-use change. Ecology Letters 10:299-314.

Kremer, R.J., Kussman, R.D. (2011) Soil quality in a pecan-kura clover alley cropping system in the Midwestern USA. Agroforestry Systems 83:213-223.

Krewenka, K.M., Holzschuh, A., Tscharntke, T., Dormann, C.F. (2011) Landscape elements of potential barriers and corridors for bees, wasps and parasitoids. Biological Conservation 144:1816-1825.

Kummerow, J. (1983) Comparative phenology of Mediterranean-type plant communities. In: F.J. Kruger, D.T. Mitchell, J.U.M. Jarvis (eds.) Mediterranean-type ecosystems, pp. 300-317. Springer-Verlag, Berlin Heidelberg.

Ladizinsky, G. (1999) On the origin of almond. Genetic resources and crop evolution46:143-147.

Lander, T.A., Bebber, D.P., Choy, C.T.L., Harris, S.A., Boshier, D.H. (2011) The Circe Principle explains how resource-rich land can waylay pollinators in fragmented landscapes. Current Biology 21:1302-1307.

Landis, D.A., Wratten, S.D., Gurr, G.M. (2000) Habitat management to conserve natural enemies of arthropod pests in agriculture. Annual Review of Entomology 45:175-201.

Larson, B.M.H., Kevan, P.G., Inouye, D.W. (2001) Flies and flowers: taxonomic diversity of anthophiles and pollinators. The Canadian Entomologist 133:439-465.

Laubertie, E.A., Wratten, S.D., Sedcole, J.R. (2006) The role of odour and visual cues in the pan-trap catching of hoverflies (Diptera: Syrphidae). Annals of Applied Biology 148:173-178.

Lavorel, S., Grigulis, K., Lamarque, P., Colace, M-P., Garden, D., Girel, J., Pellet, G., Douzet, R. (2011) Using plant functional traits to understand the landscape distribution of multiple ecosystem services. Journal of Ecology 99:135-147.

Leather, S.R., Walters, K.F.A., Bale, J.S. (1993) The Ecology of Insect Overwintering. Cambridge University Press, Cambridge.

Lebuhn, G., Droege, S., Connor, E.F., Gemmill-Herren, B., Potts, S.G., Minckley, R.L., Griswold, T., Jean, R., Kula, E., Roubik, D.W., Cane, J., Wright, K.W., Frankie, G., Parker, F. (2013) Detecting insect pollinator declines on regional and global scales. Conservation Biology 27:113-120.

147

Legendre, P., Legendre, L. (1998) Numerical Ecology. 2nd English Edn. Elsevier, Amsterdam.

Leinster, T., Cobbold, C.A. (2012) Measuring diversity: the importance of species similarity. Ecology 93:477-489.

Lentini, P.E., Martin, T.G., Gibbons, P., Fischer, J., Cunningham, S.A. (2012) Supporting wild pollinators in a temperate agricultural landscape: Maintaining mosaics of natural features and production. Biological Conservation 149:84-92.

Leponce, M., Theunis, L., Delabie, J.H.C., Roisin, Y. (2004) Scale dependence of diversity measures in a leaf-litter ant assemblage. Ecography 27:253-267.

Letourneau, D.K., Bothwell, S.G. (2008) Comparison of organic and conventional farms: challenging ecologists to make biodiversity functional. Frontiers in Ecology and Environment 6:430-438.

Letourneau, D.K., Armbrecht, I., Rivera, B., Lerma, J.M., Carmona, E.J., Daza, M.C., Escobar, S., Galindo, V., Gutiérrez, C., López, S.D., Mejía, J.L., Rangel, A.M.A., Rangel, J.H., Rivera, L., Saavedra, C.A., Torres, A.B., Trujillo, A.R. (2011) Does plant diversity benefit agroecosystems? A synthetic review. Ecological Applications 21:9-21.

Letourneau, D.K., Bothwell Allen, S.G., Stireman III, J.O. (2012) Perennial habitat fragments, parasitoid diversity nd parasitism in ephemeral crops. Journal of Applied Ecology 49:1405-1416.

Levin, S.A. (1992) The problem of pattern and scale in ecology: the Robert H MacArthur Award lecture. Ecology 73:1943-1967.

Lindenmayer, D.B., Hobbs, R.J. (2004) Fauna conservation in Australian plantation forests – a review. Biological Conservation 119:151-168.

Liu, C., Whittaker, R.J., Ma, K., Malcolm, J.R. (2007) Unifying and distinguishing diversity ordering methods for comparing communities. Population Ecology49:89-100.

Loreau, M., Naeem, S., Inchausti, P., Bengtsson, J., Grime, J.P., Hector, A., Hooper, D.U., Huston, M.A., Raffaelli, D., Schmid, B., Tilman, D., Wardle, D.A. (2001) Biodiversity and ecosystem functioning: Current knowledge and future challenges. Science 294:804-808.

Losey, J.E., Vaughan, M. (2006) The economic value of ecological services provided by insects. BioScience 56:311-323.

Lovell, S., Hamer, M., Slotow, R., Herbert, D. (2007) Assessment of congruency across invertebrate taxa and taxonomic levels to identify potential surrogates. Biological Conservation 139:113-125.

148

Lowenstein, D.M., Huseth, A.S., Groves, R.L. (2012) Response of wild bees (Hymenoptera: Apoidea: Anthophila) to surrounding land cover in Wisconsin pickling cucumber. Environmental Entomology 41:532-540.

Lucey, J.M., Hill, J.K. (2012) Spillover of insects from rain forest into adjacent oil palm plantations. Biotropica 44:368-377.

Luck, G.W., Possingham, H.P., Paton, D.C. (1999) Bird responses at inherent and induced edges in the Murray Mallee, South Australia. 1. Differences in abundance and diversity. Emu 99:157-169.

Luck, G.W., Harrington, R., Harrison, P.A., Kremen, C., Berry, P.M., Bugter, R., Dawson, T.P., de Bello, F., Díaz, S., Feld, C.K., Haslett, J.R., Hering, D., Kontogianni, A., Lavorel, S., Rounsevell, M., Samways, M.J., Sandin, L., Settele, J., Sykes, M.T., van den Hove, S., Vandewalle, M., Zobel, M. (2009) Quantifying the contribution of organisms to the provision of ecosystem services. BioScience 59:223-235.

Luck, G.W., Triplett, S., Spooner, P.G. (2013) Bird use of almond plantations: implications for conservation and production. Wildlife Research 40:523-535.

Ludwig, J.A., Tongway, D.J. (1995) Spatial organisation of landscapes and its function in semi-arid woodlands, Australia. Landscape Ecology 10:51-63.

Lundin, O., Smith, H.G., Rundlöf, M., Bommarco, R. (2013) When ecosystem services interact: crop pollination benefits depend on the level of pest control. Proceedings of the Royal Society B 280:20122243.

MacArthur, R.H., Wilson, E.O. (1963) An equilibrium theory of insular zoogeography. Evolution 17:373-387.

Macfadyen, S., Cunningham, S.A., Costamagna, A.C., Schellhorn, N.A. (2012) Managing ecosystem services and biodiversity conservation in agricultural landscapes: are the solutions the same? Journal of Applied Ecology 49:690-694.

Macfadyen, S., Muller, W. (2013) Edges in agricultural landscapes: species interactions and movement of natural enemies. PLOS One 8(3):e59659. doi:10.1371/journal.pone.0059659.

Maestre, F.T., Castillo-Monroy, A.P., Bowker, M.A., Ochoa-Hueso, R. (2012) Species richness effects on ecosystem multifunctionality depend on evenness, composition and spatial pattern. Journal of Ecology 100:317-330.

Magurran, A.E. (2004) Measuring Biological Diversity. Blackwell Science, Malden, MA.

Majer, J.D., Recher, H.F., Wellington, A.B., Woinarski, J.C.Z., Yen, A.L. (1997) Invertebrates of eucalypt forests. In: Eucalypt Ecology: Individuals to Ecosystems, pp.278-302. Eds. J. Williams, J.C.Z. Woinarski. Cambridge University Press, Cambridge.

149

Mallee Catchment Management Authority (2008) Mallee Native Vegetation Plan. Mallee Catchment Management Authority, Irymple, Victoria. Available from: http://www.malleecma.vic.gov.au.

Mandelik, Y., Roll, U. (2009) Diversity patterns of wild bees in almond orchards and their surrounding landscape. Israel Journal of Plant Sciences 57:185-191.

Mandelik, Y., Winfree, R., Neeson, T., Kremen, C. (2012) Complementary habitat use by wild bees in agro-natural landscapes. Ecological Applications 22:1535-1546.

Marini, L., Quaranta, M., Fontana, P., Biesmeijer, J.C., Bommarco, R. (2012) Landscape context and elevation affect pollinator communities in intensive apple orchards. Basic and Applied Ecology 13:681-689.

Marino, P.C., Landis, D.A. (1996) Effect of landscape structure on parasitoid diversity and parasitism in agroecosystems. Ecological Applications 6:276-284.

Markwell, T.J., Kelly, D., Duncan, K.W. (1993) Competition between honey bees (Apis mellifera) and wasps (Vespula spp.) in honeydew beech (Nothofagus solandri var. solandri) forest. New Zealand Journal of Ecology 17:85-93.

Marshall, J.W., Foot, M.A. (1979) Parasite problems of plant monoculture. New Zealand Journal of Zoology 6:656.

Mason, J. (2003) Sustainable Agriculture. Landlinks Press, Collingwood, VIC.

Matson, P.A., Parton, W.J., Power, A.G., Swift, M.J. (1997) Agricultural intensification and ecosystem properties. Science 277:504-509.

Mayer, C., Adler, L., Armbruster, W.C., Dafni, A., Eardley, C., Huang, S-Q., Kevan, P.G., Ollerton, J., Packer, L., Ssymank, A., Stout, J.C., Potts, S.G. (2011) Pollination ecology in the 21st century: key questions for future research. Journal of Pollination Ecology 3:8-23.

Mayfield, M.M., Bonser, S.P., Morgan, J.W., Aubin, I., McNamara, S., Vesk, P.A. (2010) What does species richness tell us about functional trait diversity? Predictions and evidence for responses of species and functional trait diversity to land-use change. Global Ecology and Biogeography 19:423-431.

McArdle, B.H., Anderson, M.J. (2004) Variance heterogeneity, transformations, and models of species abundance: a cautionary tale. Canadian Journal of Fisheries and Aquatic Sciences 61:1294-1302.

McKenna, M.K.L., Murray, W.E. (2002) Jungle law in the orchard: comparing globalisation in the New Zealand and Chilean apple industries. Economic Geography 78:494-514.

Melo, A.S., Pereira, R.A.S., Santos, A.J., Shepherd, G.J., Machado, G., Medeiros, H.F., Sawaya, R.J. (2003) Comparing species richness among assemblages using

150

sample units: why not use extrapolation methods to standardise different sample sizes? Oikos 101:398-410.

Meng, L-Z., Martin, K., Liu, J-X., Burger, F., Chen, J. (2012) Contrasting responses of hoverflies and wild bees to habitat structure and land use change in a tropical landscape (southern Yunnan, SW China). Insect Science 19:666-676.

Menhinick, E.F. (1964) A comparison of some species-individuals diversity indices applied to samples of field insects. Ecology 45:859-861.

Mesa, L.A., Howlett, B.G., Grant, J.E., Didham, R.K. (2013) Changes in the relative abundance and movement of insect pollinators during the flowering cycle of Brassica rapa crops: implications for gene flow. Journal of Insect Science 13:13.

Michener, C.D. (1979) Biogeography of the bees. Annals of the Missouri Botanical Garden 66:277-347.

Michener, C.D. (2007) The Bees of the World. Johns Hopkins University Press, Baltimore MD.

Morton, S.R., Stafford Smith, D.M., Friedel, M.H., Griffin, G.F., Pickup, G. (1995) The stewardship of arid Australia: ecology and landscape management. Journal of Environmental Management 43:195-217.

Muff, P., Kropf, C., Frick, H., Nentwig, W., Schmidt-Entling, M.H. (2009) Co-existence of divergent communities at natural boundaries: spider (Arachnida: Araneae) diversity across an alpine timberline. Insect Conservation and Diversity 2:36-44.

Murray, T.E., Fitzpatrick, U., Byrne, A., Fealy, R., Brown, M.J.F., Paxton, R.J. (2012) Local-scale factors structure wild bee communities in protected areas. Journal of Applied Ecology 49:998-1008.

Naveh, Z., Whittaker, R.H. (1979) Structural and floristic diversity of shrublands and woodlands in northern Israel and other Mediterranean areas. Vegetatio 41:171-190.

Nekola, J.C., White, P.S. (1999) The distance decay of similarity in biogeography and ecology. Journal of Biogeography 26:867-878.

New, T.R., Yen, A.L., Sands, D.P.A., Greenslade, P., Neville, P.J., York, A., Collett, N.G. (2010) Planned fires and invertebrate conservation in south east Australia. Journal of Insect Conservation 14:567-574.

Nicholls, C.I., Altieri, M.A. (2013) Plant biodiversity enhances bees and other insect pollinators in agroecosystems. A review. Agronomy for Sustainable Development33:257-274.

Nielsen, A., Dauber, J., Kunin, W.E., Lamborn, E., Jauker, B., Moora, M., Potts, S.G., Reitan, T., Roberts, S., Sõber, V., Settele, J., Steffan-Dewenter, I., Stout, J.C., Tscheulin, T., Vaitis, M., Vivarelli, D., Biesmeijer, J.C., Petanidou, T. (2012)

151

Pollinator community responses to the spatial population structure of wild plants: a pan-European approach. Basic and Applied Ecology 13:489-499.

Noble, J.C., Bradstock, R.A. (eds.) (1989) Mediterranean Landscapes in Australia: Mallee Ecosystems and their Management. CSIRO, Melbourne VIC.

NSW Agriculture (2002) Irrigation for Horticulture in the Mallee. NSW Agriculture, Orange NSW.

Obrist, M.K., Duelli, P. (2010) Rapid biodiversity assessment of arthropods for monitoring average local species richness and related ecosystem services. Biodiversity and Conservation 19:2201-2220.

Öckinger, E., Smith, H.G. (2007) Semi-natural grasslands as population sources for pollinating insects in agricultural landscapes. Journal of Applied Ecology 44:50-59.

Odum, E.P. (1969) The strategy of ecosystem development. Science 164:262-270.

O’Hara, R.B., Kotze, D.J. (2010) Do not log-transform count data. Methods in Ecology and Evolution 1:118-122.

Oldroyd, B.P., Lawler, S.H., Crozier, R.H. (1994) Do feral honey bees (Apis mellifera) and regent parrots (Polytelis anthopeplus) compete for nest sites? Australian Journal of Ecology 19:444-450.

Oliver, I., Beattie, A.J. (1996) Invertebrate morphospecies as surrogates for species: a case study. Conservation Biology 10:99-109.

Olson, D., Andow, D. (2008) Patch edges and insect populations. Oecologia 155:549-558.

Onans, J., Parsons, R.F. (1980) Regeneration of native plants on abandoned mallee farmland in south-eastern Australia. Australian Journal of Botany 28:479-493.

Osterman, L.E., Buzas, M.A., Hayek, L-A.C. (2002) SHE analysis for biozonation of benthic foraminiferal assemblages from western Arctic Ocean. Palaios 17:297-303.

Ostfeld, R.S., Keesing, F. (2000) Pulsed resources and community dynamics of consumers in terrestrial ecosystems. Trends in Ecology and Evolution 15:232-237.

Paini, D.R., Williams, M.R., Roberts, J.D. (2005) No short-term impact of honey bees on the reproductive success of an Australian native bee. Apidologie 36:613-621.

Pal, R.W., Pinke, G., Botta-Dukát, Z., Campetella, G., Bartha, S., Kalocsai, R., Lengyel, A. (2013) Can management intensity be more important than environmental factors? A case study along an extreme elevation gradient from central Italian cereal fields. Plant Biosystems – An International Journal Dealing with all

152

Aspects of Plant Biology: Official Journal of the Societa Botanica Italiana147:343-353.

Paredes, D., Cayuela, L., Campos, M. (2013) Synergistic effects of ground cover and adjacent vegetation on natural enemies of olive insect pests. Agriculture, Ecosystems and Environment 173:72-80.

Parra-H, A., Nates-Parra, G. (2012) The ecological basis for biogeographic classification: an example in orchid bees (Apidae: Euglossini). Neotropical Entomology 41:442-449.

Patil, G.P., Taillie, C. (1982) Diversity as a concept and its measurement. Journal of the American Statistical Association 77:548-561.

Pauw, A. (2007) Collapse of a pollination web in small conservation areas. Ecology88:1759-1769.

Pauw, A., Hawkins, J.A. (2011) Reconstruction of historical pollination rates reveals linked declines of pollinators and plants. Oikos 120:344-349.

Peakall, R., Handel, S.N. (1993) Pollinators discriminate among floral heights of a sexually deceptive orchid: implications for selection. Evolution 47:1681-1687.

Persson, A.S., Smith, H.G. (2013) Seasonal persistence of bumblebee populations is affected by landscape context. Agriculture, Ecosystems and Environment165:201-209.

Petanidou, T., Vokou, D. (1990) Pollination and pollen energetic in Mediterranean ecosystems. American Journal of Botany 77:986-992.

Petanidou, T., Lamborn, E. (2008) A land for flowers and bees: studying pollination ecology in Mediterranean communities. Plant Biosystems - An International Journal Dealing with all Aspects of Plant Biology: Official Journal of the Societa Botanica Italiana 139:279-294.

Peters, V.E., Carroll, C.R., Cooper, R.J., Greenberg, R., Solis, M. (2013) The contribution of plant species with a steady-state flowering phenology to native bee conservation and bee pollination services. Insect Conservation and Diversity 6:45-56.

Pfiffner, L., Luka, H. (2000) Overwintering of arthropods in soils of arable fields and adjacent semi-natural habitats. Agriculture, Ecosystems and Environment 78:215-222.

Pickering, C.M., Stock, M. (2003) Insect colour preference compared to flower colours in the Australian Alps. Nordic Journal of Botany 23:217-223.

Pik, A.J., Oliver, I., Beattie, A.J. (1999) Taxonomic sufficiency in ecological studies of terrestrial invertebrates. Australian Journal of Ecology 24:555-562.

153

Pimentel, D., Culliney, T.W., Buttler, I.W., Reinemann, D.J., Beckman, K.B. (1989) Low-input sustainable agriculture using ecological management practices. Agriculture, Ecosystems and Environment 27:3-24.

Pinke, G., Pal, R. (2009) Floristic composition and conservation value of the stubble-field weed community, dominated by Stachys annua in western Hungary. Biologia 64:279-291.

Plourde, J.D., Pijanowski, B.C., Pekin, B.K. (2013) Evidence for increased monoculture cropping in the Central United States. Agriculture, Ecosystems and Environment165:50-59.

Popic, T.J., Davila, Y.C., Wardle, G.M. (2013) Evaluation of common methods for sampling invertebrate pollinator assemblages: net sampling out-perform pan traps. PLoS ONE 8(6):e66665. doi:10.1371/journal.pone.0066665.

Potts, S.G., Vulliamy, B., Dafni, A., Ne’eman, G., Willmer, P. (2003) Linking bees and flowers: how do floral communities structure pollinator communities? Ecology84:2628-2642.

Potts, S.G., Vulliamy, B., Roberts, S., O’Toole, C., Dafni, A., Ne’eman, G., Willmer, P. (2005) Role of nesting resources in organising diverse bee communities in a Mediterranean landscape. Ecological Entomology 30:78-85.

Potts, S.G., Biesmeijer, J.C., Kremen, C., Neumann, P., Schweiger, O., Kunin, W.E. (2010) Global pollinator declines: trends, impacts and drivers. Trends in Ecology and Evolution 25:345-353.

Preston, F.W. (1960) Time and space and the variation of species. Ecology 41:612-627.

Proctor, E., Nol, E., Burke, D., Crins, W.J. (2012) Responses of insect pollinators and understory plants to silviculture in northern hardwood forests. Biodiversity and Conservation 21:1703-1740.

Pryke, J.S., Samways, M.J. (2012) Conservation management of complex natural forest and plantation edge effects. Landscape Ecology 27:73-85.

Pucci, T. (2008) A comparison of the parasitic wasps (Hymenoptera) at elevated versus ground yellow pan traps in a beech-maple forest. Journal of Hymenoptera Research 17:116-123.

Pywell, R.F., Meek, W.R., Hulmes, L., Hulmes, S., James, K.L., Nowakowski, M., Carvell, C. (2011) Management to enhance pollen and nectar resources for bumblebees and butterflies within intensively farmed landscapes. Journal of Insect Conservation 15:853-864.

Quine, C.P., Humphrey, J.W. (2010) Plantations of exotic tree species in Britain: irrelevant for biodiversity or novel habitat for native species? Biodiversity and Conservation 19:1503-1512.

154

Rader, R., Edwards, W., Westcott, D.A., Cunningham, S.A., Howlett, B.G. (2011) Pollen transport differs among bees and flies in a human-modified landscape. Diversity and Distributions 17:519-529.

Rader, R., Howlett, B.G., Cunningham, S.A., Westcott, D.A., Edwards, W. (2012) Spatial and temporal variation in pollinator effectiveness: do unmanaged insects provide consistent pollination services to mass flowering crops? Journal of Applied Ecology 49:126-134.

Rader, R., Edwards, W., Westcott, D.A., Cunningham, S.A., Howlett, B.G. (2013) Diurnal effectiveness of pollination by bees and flies in agricultural Brassica rapa: implications for ecosystem resilience. Basic and Applied Ecology 14:20-27.

Ramirez, P.A., Simonetti, J.A. (2011) Conservation opportunities in commercial plantations: the case of mammals. Journal for Nature Conservation 19:351-355.

Ramos, M.E., Benítez, E., García, P.A., Robles, A.B. (2010) Cover crops under different management vs. frequent tillage in almond orchards in semiarid conditions: effects on soil quality. Applied Soil Ecology 44:6-14.

Rand, T.A., Tylianakis, J.M., Tscharntke, T. (2006) Spillover edge effects: the dispersal of agriculturally subsidized insect natural enemies into adjacent natural habitats. Ecology Letters 9:603-614.

Randlkofer, B., Obermaier, E., Hilker, M., Meiners, T. (2010a) Vegetation complexity –the influence of plant species diversity and plant structures on plant chemical complexity and arthropods. Basic and Applied Ecology 11:383-395.

Randlkofer, B., Obermaier, E., Casas, J., Meiners, T. (2010b) Connectivity counts: disentangling effects of vegetation structure elements on the searching movement of a parasitoid. Ecological Entomology 35:446-455.

Rattigan, K., Hill, S.J. (1987) Relationship between temperature and flowering in almond – effect of location. Australian Journal of Experimental Agriculture 27:905-908.

Raven, P.H., Yeates, D.K. (2007) Australian biodiversity: threats for the present, opportunities for the future. Australian Journal of Entomology 46:177-187.

Reese, K.M., Philpott, S.M. (2010) Environmental and habitat drivers of relative abundance for a suite of Azteca-attacking Pseudacteaon phorid flies. Environmental Entomology 41:1107-1114.

Ricketts, T.H., Regetz, J., Steffan-Dewenter, I., Cunningham, S.A., Kremen, C., Bogdanski, A., Gemmill-Herren, B., Greenleaf, S.S., Klein, A-M., Mayfield, M.M., Morandin, M.A., Ochieng, A., Viana, B.F. (2008) Landscape effects on crop pollination services: are there general patterns? Ecology Letters 11:499-515.

Ries, L., Fletcher Jr, R.J., Battin, J., Sisk, T.D. (2004) Ecological responses to habitat edges: mechanisms, models and variability explained. Annual Review of Ecology, Evolution and Systematics 35:491-522.

155

Robinson, S.L., Lundholm, J.T. (2012) Ecosystem services provided by urban spontaneous vegetation. Urban Ecosystems 15:545-557.

Robson, D.B. (2013) An assessment of the potential for pollination facilitation of a rare plant by common plants: Symphyotrichum sericeum (Asteraceae) as a case study. Botany 91:34-42.

Rollin, O., Bretagnolle, V., Decourtye, A., Aptel, J., Michel, N., Vaissière, B.E., Henry, M. (2013) Differences of floral resource use between honey bees and wild bees in an intensive farming system. Agriculture, Ecosystems and Environment 179:78-86.

Roschewitz, I., Gabriel, D., Tscharntke, T., Thies, C. (2005) The effects of landscape complexity on arable weed species diversity in organic and conventional farming. Journal of Applied Ecology 42:873-882.

Roubik, D.W., Wolda, H. (2001) Do competing honey bees matter? Dynamics and abundance of native bees before and after honey bee invasion. Population Ecology 43:53-62.

Ruiz-Colmenero, M., Bienes, R., Marques, M.J. (2011) Soil and water conservation dilemmas associated with the use of green cover in steep vineyards. Soil & Tillage Research 117:211-223.

Rundlöf, M., Nilsson, H., Smith, H.G. (2008) Interacting effects of farming practice and landscape context on bumble bees. Biological Conservation 141:417-426.

Ryszkowski, L., Kędziora, A. (1987) Impact of agricultural landscape structure on energy flow and water cycling. Landscape Ecology 1:85-94.

Sadras, V., Roget, D., Krause, M. (2003) Dynamic cropping strategies for risk management in dry-land farming systems. Agricultural Systems 76:929-948.

Sala, O.E., Lauenroth, W.K. (1982) Small rainfall events: an ecological role in semi-arid regions. Oecologia 53:301-304.

Samnegård, U., Persson, A.S., Smith, H.G. (2011) Gardens benefit bees and enhance pollination in intensively managed farmland. Biological Conservation 144:2602-2606.

Saunders, M.E., Luck, G.W. (2013) Pan trap catches of pollinator insects vary with habitat. Australian Journal of Entomology 52:106-113.

Saunders, M.E., Luck, G.W., Mayfield, M.M. (2013) Almond orchards with living ground cover host more wild insect pollinators. Journal of Insect Conservation17:1011-1025.

Schiegg, K. (2000) Effects of dead wood volume and connectivity on saproxylic insect species diversity. Ecoscience 7:290-298.

156

Seavy, N.E., Quader, S., Alexander, J.D., Ralph, C.J. (2005) Generalised Linear Models and point count data: statistical considerations for the design and analysis of monitoring studies. USDA Forest Service General Technical Report, PSW-GTR-191:744-753. URL: http://www.fs.fed.us/psw/publications/documents/psw_gtr191/psw_gtr191_0744-0753_seavy.pdf. Accessed 24 February 2012.

Şekercioğlu, C.H. (2010) Ecosystem functions and services. In: N.S. Sodhi, P.R. Ehrlich (eds.) Conservation Biology for All, pp. 45-72. Oxford University Press, Oxford.

Sharley, D.J., Hoffmann, A.A., Thomson, L.J. (2008) The effects of soil tillage on beneficial invertebrates within the vineyard. Agricultural and Forest Entomology10:233-243.

Sheffield, C.S., Westby, S.M., Smith, R.F., Kevan, P.G. (2008) Potential of bigleaf lupine for building and sustaining Osmia lignaria populations for pollination of apple. The Canadian Entomologist 140:589-599.

Sileshi, G. (2006) Selecting the right statistical model for analysis of insect count data by using information theoretic measures. Bulletin of Entomological Research 96:479-488.

Singleton, G., Krebs, C.J., Davis, S., Chambers, L., Brown, P. (2001) Reproductive changes in fluctuating house mouse populations in southeastern Australia. Proceedings of the Royal Society of London B 268:1741-1748.

Small, C.J., McCarthy, B.C. (2002) Spatial and temporal variability of herbaceous vegetation in an eastern deciduous forest. Plant Ecology 164:37-48.

Smith, J.R. (1916) The oak tree and man’s environment. Geographical Review 1:3-19.

Smith, F.P., Prober, S.M., House, A.P.N., McIntyre, S. (2013) Maximising retention of native biodiversity in Australian agricultural landscapes – the 10:20:40:30 guidelines. Agriculture, Ecosystems and Environment 166:35-45.

Soininen, J. (2010) Species turnover along abiotic and biotic gradients: patterns in space equal patterns in time? BioScience 60:433-439.

Spivak, M., Mader, E., Vaughan, M., Euliss Jr, N.H. (2011) The plight of the bees. Environmental Science & Technology 45:34-38.

Srivastava, D.S. (2006) Habitat structure, trophic structure and ecosystem function: interactive effects in a bromeliad-insect community. Oecologia 149:493-504.

Ssymank, A., Kearns, C.A., Pape, T., Thompson, F.C. (2008) Pollinating flies (Diptera): a major contribution to plant diversity and agricultural production. Biodiversity9:86-89.

157

Stasek, D.J., Bean, C., Crist, T.O. (2008) Butterfly abundance and movements among prairie patches: the role of habitat quality, edge, and forest matrix permeability. Environmental Entomology 37:897-906.

Steffan-Dewenter, I., Münzenberg, U., Tscharntke, T. (2001) Pollination, seed set and seed predation on a landscape scale. Proceedings of the Royal Society of London B 268:1685-1690.

Steffan-Dewenter, I., Münzenberg, U., Bürger, C., Thies, C., Tscharntke, T. (2002) Scale-dependent effects of landscape context on three pollinator guilds. Ecology83:1421-1432.

Steffan-Dewenter, I., Leschke, K. (2003) Effects of habitat management on vegetation and above-ground nesting bees and wasps of orchard meadows in Central Europe. Biodiversity and Conservation 12:1953-1968.

Stephens, C.J., Schellhorn, N.A., Wood, G.M., Austin, A.D. (2006) Parasitic wasp assemblages associated with native and weedy plant species in an agricultural landscape. Australian Journal of Entomology 45:176-184.

Stevens, N.B., Stephens, C.J., Iqbal, M., Jennings, J.T., La Salle, J., Austin, A.D. (2007) What wasp is that?: an interactive identification guide to the Australasian families of Hymenoptera (CD-ROM). Australian Biological Resources Study, Canberra, ACT.

Stoate, C., Báldi, A., Boatman, N.D., Herzon, I., van Doorn, A., de Snoo, G.R., Rakosy, L., Ramwell, C. (2009) Ecological impacts of early 21st century agricultural change in Europe – a review. Journal of Environmental Management 91:22-46.

Stoddard, F.L., Bond, D.A. (1987) The pollination requirements of the faba bean. Bee World 68:144-152.

SunRISE 21 Inc. (2008) Irrigated Horticulture of the Lower Murray-Darling 1997-2006. SunRISE 21, Mildura VIC. URL: http://www.sunrise21.org.au/page.php?id=34&rid=59. Accessed September 21 2011.

Symstad, A.J., Chapin III, F.S., Wall, D.H., Gross, K.L., Huenneke, L.F., Mittelbach, G.G., Peters, D.P.C., Tilman, D. (2003) Long-term and large-scale perspectives on the relationship between biodiversity and ecosystem functioning. BioScience53: 89-98.

Tedders, W.L. (1983) Insect management in deciduous orchard ecosystems: habitat manipulation. Environmental Management 7:29-34.

Tews, J., Brose, U., Grimm, V., Tielbörger, K., Wichmann, M.C., Schwager, M., Jeltsch, F. (2004) Animal species diversity driven by heterogeneity/diversity: the importance of keystone structures. Journal of Biogeography 31:79-92.

158

Thevathasan, N.V., Gordon, A.M. (2004) Ecology of tree intercropping systems in the North temperate region: Experiences from southern Ontario, Canada. Agroforestry Systems 61-2:257-268.

Thomson, J.D. (1981) Spatial and temporal components of resource assessment by flower-feeding insects. Journal of Animal Ecology 50:49-59.

Thomson, J.D., Goodell, K. (2001) Pollen removal and deposition by honeybee and bumblebee visitors to apple and almond flowers. Journal of Applied Ecology38:1032-1044.

Tilman, D. (1999) Global environmental impacts of agricultural expansion: the need for sustainable and efficient practices. Proceedings of the National Acadamey of Sciences USA 96:5995-6000.

Tóthmérész, B. (1995) Comparison of different methods for diversity ordering. Journal of Vegetation Science 6:283-290.

Tscharntke, T., Steffan-Dewenter, I., Kruess, A., Thies, C. (2002) Characteristics of insect populations on habitat fragments: a mini review. Ecological Research17:229-239.

Tscharntke, T., Klein, A-M., Kruess, A., Steffan-Dewenter, I., Thies, C. (2005) Landscape perspectives on agricultural intensification and biodiversity –ecosystem service management. Ecology Letters 8:857-874.

Tscharntke, T., Tylianakis, J.M., Rand, T.A., Didham, R.K., Fahrig, L., Batáry, P., Bengtsson, J., Clough, Y., Crist, T.O., Dormann, C.F., Ewers, R.M., Fründ, J., Holt, R.D., Holzschuh, A., Klein, A-M., Kleijn, D., Kremen, C., Landis, D.A., Laurance, W., Lindenmayer, D., Scherber, C., Sodhi, N., Steffan-Dewenter, I., Thies, C., van der Putten, W.H., Westphal, C. (2012) Landscape moderation of biodiversity patterns and processes – eight hypotheses. Biological Reviews87:661-685.

Tscheulin, T., Neokosmidis, L., Petanidou, T., Settele, J. (2011) Influence of landscape context on the abundance and diversity of bees in Mediterranean olive groves. Bulletin of Entomological Research 101:557-564.

Tuell, J.K., Isaacs, R. (2009) Elevated pan traps to monitor bees in flowering crop canopies. Entomologia Experimentalis et Applicata 131:93-98.

Tuomisto, H., Ruokolainen, K. (2006) Analysing or explaining beta diversity? Understanding the targets of different methods of analysis. Ecology 87:2697-2708.

Turner, M.G. (1989) Landscape ecology: The effect of pattern on process. Annual Review of Ecology, Evolution and Systematics 20:171-197.

Ulanowicz, R.E. (2002) The balance between adaptability and adaptation. BioSystems64:13-22.

159

van der Maarel, E. (1990) Ecotones and ecoclines are different. Journal of Vegetation Science 1:135-138.

Vandermeer, J., Perfecto, I., Philpott, S. (2010) Ecological complexity and pest control in organic coffee production: uncovering an autonomous ecosystem service. BioScience 60:527-537.

vanEngelsdorp, D., Hayes Jr, J., Underwood, R.M., Pettis, J. (2008) A Survey of Honey Bee Colony Losses in the U.S., Fall 2007 to Spring 2008. PLoS ONE 3:e4071. doi:10.1371/journal.pone.0004071.

Vasseur, C., Joannon, A., Aviron, S., Burel, F., Meynard, J-M., Baudry, J. (2013) The cropping systems mosaic: how does hidden heterogeneity of agricultural landscapes drive arthropod populations? Agriculture, Ecosystems and Environment 166:3-14.

Vezvaei, A. (1994) Pollination Studies in Almond. PhD Thesis, University of Adelaide.

Vezvaei, A., Jackson, J.F. (1995) Effect of pollen parent and stages of flower development on almond nut production. Australian Journal of Experimental Agriculture 35:109-113.

Vogt, W. (1948) Road to Survival. William Sloan, New York NY.

Vrdoljak, S.M., Samways, M.J. (2012) Optimising coloured pan traps to survey flower visiting insects. Journal of Insect Conservation 16:345-354.

Wäckers, F.L., van Rijn, P.C.J., Bruin, J. Eds. (2005) Plant-provided food for carnivorous insects: a protective mutualism and its applications. Cambridge University Press, Cambridge UK.

Walton, N.J., Isaacs, R. (2011) Influence of native flowering plant strips on natural enemies and herbivores in adjacent blueberry fields. Environmental Entomology40:697-705.

Warren, M.S., Key, R.S. (1991) Woodlands: Past, present and potential for insects. In: N.M. Collins (ed.) The Conservation of Insects and Their Habitats, pp. 155-212. Academic Press Ltd, London.

Warton, D.I. (2008) Raw data graphing: an informative but under-utilised tool for the analysis of multivariate abundances. Austral Ecology 33:290-300.

Warton, D.I., Hui, F.K.C. (2011) The arcsine is asinine: the analysis of proportions in ecology. Ecology 92:3-10

Warton, D.I., Wright, S.T., Wang, Y. (2012) Distance-based multivariate analyses confound location and dispersion effects. Methods in Ecology and Evolution 3:89-101.

160

Watson, J.C., Wolf, A.T., Ascher, J.S. (2011) Forested landscapes promote richness and abundance of native bees (Hymenoptera: Apoidea: Anthophila) in Wisconsin apple orchards. Environmental Entomology 40:621-632.

Weinbaum, S.A. (1985) Role of natural self-pollination in self-fruitfulness of almond. Scientia Horticulturae 27:295-302.

Weis, T. (2010) The accelerating biophysical contradictions of industrial capitalist agriculture. Journal of Agrarian Change 10:315-341.

Westgarth, W. (1848) Australia Felix; or, A historical and descriptive account of the settlement of Port Phillip, New South Wales. Oliver & Boyd, Edinburgh.

Westphal, C., Steffan-Dewenter, I., Tscharntke, T. (2003) Mass flowering crops enhance pollinator densities at a landscape scale. Ecology Letters 6:961-965.

Westphal, C., Bommarco, R., Carré, G., Lamborn, E., Morison, N., Petanidou, T., Potts, S.G., Roberts, S.P.M., Szentgyörgyi, H., Tscheulin, T., Vaissière, B., Woyciechowski, M., Biesmeijer, J.C., Kunin, W.E., Settele, J., Steffan-Dewenter, I. (2008) Measuring bee diversity in different European habitats and biogeographical regions. Ecological Monographs 78:653-671.

Westphal, C., Steffan-Dewenter, I., Tscharntke, T. (2009) Mass flowering oilseed rape improves early colony growth but not sexual reproduction of bumblebees. Journal of Applied Ecology 46:187-193.

Westrich, P. (1996) Habitat requirements of central European bees and the problem of partial habitats. In: A. Matheson, S.L. Buchmann, C. O'Toole, P. Westrich, I.H. Williams (eds.) The conservation of bees, pp. 1-16. Academic Press, New York NY.

Whitford, W.G. (1996) The importance of the biodiversity of soil biota in arid ecosystems. Biodiversity and Conservation 5:185-195.

Whittaker, R.H., Niering, W.A., Crisp, M.D. (1979) Structure, pattern and diversity of a mallee community in New South Wales. Vegetatio 39:65-76.

Wilkinson, J. (2012) Almonds in Australia: from pioneer planting to prime production. Almond Board of Australia, Berri SA.

Williams, N.M., Kremen, C. (2007) Resource distributions among habitats determine solitary bee offspring production in a mosaic landscape. Ecological Applications17:910-921.

Wilson, B. (2008) Using Shebi (SHE Analysis for Biozone Identification): to proceed from the top down or the bottom up? A discussion using two Miocene foraminiferal successions from Trinidad, West Indies. Palaios 23:636-644.

Winfree, R., Griswold, T., Kremen, C. (2007) Effect of human disturbance on bee communities in a forested ecosystem. Conservation Biology 21:213-223.

161

Winfree, R., Williams, N.M., Gaines, H., Ascher, J.S., Kremen, C. (2008) Wild bee pollinators provide the majority of crop visitation across land-use gradients in New Jersey and Pennsylvania, USA. Journal of Applied Ecology 45:793-802.

Winfree, R., Bartomeus, I., Cariveau, D.P. (2011) Native pollinators in anthropogenic habitats. Annual Review of Ecology, Evolution and Systematics 42:1-22.

Winqvist, C., Ahnström, J., Bengtsson, J. (2012) Effects of organic farming on biodiversity and ecosystem services: taking landscape complexity into account. Annals of the New York Academy of Sciences 1249:191-203.

With, K.A., Pavuk, D.M. (2011) Habitat area trumps fragmentation effects on arthropods in an experimental landscape system. Landscape Ecology 26:1035-1048.

Wood, J.G. (1929) Floristics and ecology of the mallee. Transactions of the Royal Society of South Australia 53:359-78.

Woodhouse, P. (2010) Beyond industrial agriculture? Some questions about farm size, productivity and sustainability. Journal of Agrarian Change 10:437-453.

Wratten, S.D., Bowie, M.H., Hickman, J.M., Evans, A.M., Sedcole, J.R., Tylianakis, J.M. (2003) Field boundaries as barriers to movement of hover flies (Diptera: Syrphidae) in cultivated land. Oecologia 134:605-611.

Wratten, S.D., Gillespie, M., Decourtye, A., Mader, E., Desneux, N. (2012) Pollinator habitat enhancement: Benefits to other ecosystem services. Agriculture, Ecosystems and Environment 159:112-122.

Yahner, R.H. (1988) Changes in wildlife communities near edges. Conservation Biology2:333-339.

Yarrow, M.M., Marín, V.H. (2007) Toward conceptual cohesiveness: a historical analysis of the theory and utility of ecological boundaries and transition zones. Ecosystems 10:462-476.

Yen, A.L., Ewart, D., Walker, K. (2006) Mallee eucalypts, hummock grasses and social insects – key elements of the Victorian mallee. Proceedings of the Royal Society of Victoria 118:281-293.

Yong, K.E., Li, Y., Hendrix, S.D. (2012) Habitat choice of multiple pollinators in almond trees and its potential effect on pollen movement and productivity: a theoretical approach using the Shigesada-Kawasaki-Teramoto model. Journal of Theoretical Biology 305:103-109.

Yu, X.D., Luo, T.H., Zhou, H.Z., Yang, J. (2007) Distribution of carabid beetles (Coleoptera: Carabidae) across a forest-grassland ecotone in southwestern China. Environmental Entomology 36:348-355.

162

Zurbuchen, A., Landert, L., Klaiber, J., Müller, A., Hein, S., Dorn, S. (2010) Maximum foraging ranges in solitary bees: only few individuals have the capability to cover long foraging distances. Biological Conservation 143:669-676.

163

APPENDICES

Appendix 1: Pan trap catches of pollinator insects vary with habitat

ABSTRACT

Coloured pan trapping is a simple and efficient method for collecting flying

insects, yet there is still discussion over the most effective bowl colour to use for

particular target groups (e.g. pollinator insects). The success of particular colours

can vary across bioregions and habitat contexts. Most published pan trap studies

have been conducted in the northern hemisphere and very few investigated the

effects of habitat context on pan trap catches. Our study is one of the first to (a)

sample for potential pollinators in Australian mallee vegetation and almond

orchards and (b) to investigate whether habitat context interacts with trap colour

to influence pan trap catches. We sampled Hymenoptera and Diptera using

yellow, white and blue pan traps in native mallee vegetation and two types of

managed almond orchards (monoculture and plant-diverse) in the Murray Mallee

bioregion of northwest Victoria, Australia. Yellow traps caught the most insects

across all habitats, although catches in each coloured trap varied with habitat

context. For all insect groups combined, blue traps caught more individuals in

mallee habitats than in almond orchards. For native hymenopterans, yellow traps

caught more individuals in plant-diverse orchards than in native sites, while blue

traps caught more individuals in native sites. Our results highlight the importance

of considering the habitat context of individual pan trapping surveys, as no one

trap colour is likely to be suitable for trapping target insects across all habitat

contexts.

*This appendix is published as Saunders ME, Luck GW (2013) Pan trap catches

of pollinator insects vary with habitat. Australian Journal of Entomology 52:106-

113. (Formatting is per journal style.)

164

INTRODUCTION

Coloured pan (or bowl) trapping is considered a simple, efficient method for

collecting flying insects, and its lack of collector bias (compared to, for example,

sweep netting) makes it particularly useful for comparing invertebrate

communities across time, locations or when collector experience varies (Cane et

al. 2000; Westphal et al. 2008). Pan trapping was traditionally used for sampling

agricultural pests and phytophagous insects (e.g. Evans & Medler 1967; Boiteau

1983), but now is also used for collecting bees (Hymenoptera: Apoidea) and other

pollinating insects (Leong & Thorp 1999; Campbell & Hanula 2007; Gollan et al.

2011).

Despite the benefits of pan trapping as a sampling method, examination of

its use for sampling pollinator insects remains limited. Potential sampling biases

associated with the technique have been identified recently (Roulston et al. 2007;

Baum & Wallen 2011) and it is not clear which bowl colour is most effective in

particular contexts (i.e. the colour that attracts the highest abundance and diversity

of the target species). Blue traps have been endorsed as the best bee attractant

(Cane et al. 2000; Stephen & Rao 2005), yet the findings of research using pan

traps to sample bees and other pollinator insects do not show a consistent

preference for a particular bowl colour (e.g. Campbell & Hanula 2007; Gollan et

al. 2011; Grundel et al. 2011).

The success of certain pan trap colours in attracting pollinators may vary

across bioregions, habitat contexts or topography. For example, Grundel et al.

(2011) sampled sites in northwest Indiana, USA, using blue, yellow and white

traps, and caught the most bees in white pan traps. Campbell and Hanula (2007)

sampled sites in south-eastern USA, including coastal plain and southern

Appalachian Mountain ecosystems, with red, blue, yellow and white pan traps and

caught the most bees in blue traps. Gollan et al. (2011) sampled sites in the New

South Wales North Coast, Sydney Basin and South Western Slopes regions in

eastern Australia, using yellow and white traps and caught the most bees in yellow

pans.

To our knowledge, there is only one published study from pan trapping

pollinator insects in Australian habitats (Gollan et al. 2011). Given the wide

variation in results from pan trap sampling in the northern hemisphere, more pan

trap studies are needed from the southern hemisphere to determine whether

165

catches in particular coloured traps are likely to vary with habitat or ecological

context. Our study is one of the first to sample potential pollinators in the Murray

Mallee region of northwest Victoria (Australia) using coloured pan traps and also

one of the first to investigate whether habitat context interacts with trap colour to

influence pan trap catches (see Missa et al. 2009; Abrahamczyk et al. 2010;

Droege et al. 2010 for relevant discussions on habitat context).

In this paper, the term ‘pollinator insect’ refers to any species of insect

belonging to the taxonomic order Hymenoptera or Diptera. Although we did not

quantify pollination efficiency in this study, we use the term to refer to flower-

visiting species of these orders so as to maintain consistency with the pollination

literature. Bees are the most recognised pollinators, yet wasps and flies are also

known to be frequent flower visitors and pollinators (Kevan & Baker 1983;

Larson et al. 2001) and many of these species are considered important, yet

understudied, generalist pollinators in a variety of natural (Pickering & Stock

2003; Brunet 2009) and crop systems (Klein et al. 2007; Jauker et al. 2009; Rader

et al. 2011). Hence, we focussed on Hymenoptera and Diptera, as the broader

scope of our research was to quantify potential native pollination services for

almond orchards that could complement current commercial European honeybee

(Apis mellifera L.) stocks. The questions addressed in the following study are:

Is there a difference in the number of pollinator insects caught in different

coloured pan traps?

Does pollinator insect abundance in each trap colour vary with habitat

context?

MATERIALS AND METHODS

Study area and system

The study area is located in northwest Victoria, Australia, in the Victorian Murray

Mallee Bioregion, specifically in the Wemen/Liparoo (hereafter Wemen) district

and the Colignan/Nangiloc (hereafter Colignan) district (see Figure 2.1). The

Murray Mallee bioregion is classified as semi-arid and annual rainfall is between

250-300 mm per year (Natural Resources & Environment n.d.). Sampling was

conducted in 2010, during which the region recorded wetter than average rainfall.

Approximately 528 mm fell in Wemen and 427 mm in Colignan. Average

166

temperatures for the months of sampling (July—September 2010) ranged from

minimums of 5—7oC to maximums of 15—20oC (Bureau of Meteorology 2011).

One natural and one agricultural habitat were sampled – native mallee

vegetation and commercial almond orchards respectively. Native vegetation sites

were located within patches of remnant mallee vegetation that were greater than

50 m across and proximate to an almond orchard. Almond orchards were

classified as either broadacre monoculture (MONO) orchards or plant-diverse

(GRASS) orchards. MONO orchards are managed intensively and have

completely bare ground across the entire orchard. They operate on a large,

broadacre scale, occupying thousands of hectares of continuous land with small,

infrequent patches of native vegetation interspersed between blocks of almond

trees. GRASS orchards are managed with lower intensity and have a living carpet

of grass and weedy ground cover across the orchard throughout the year. These

orchards are situated within a heterogeneous agricultural landscape, with smaller

block sizes (<1000 ha) interspersed among a variety of crops and land uses,

including small, frequent patches of remnant vegetation.

Figure A1.1 The study region and sampling localities in northwest Victoria.

Sampling

Insect assemblages were sampled within almond orchards and native mallee

vegetation prior to almond flowering in July 2010, and again during the final

stages of flowering in August—September 2010. Due to weather conditions and

167

logistics, the first sampling period was 5 days and the second was 10 days. Insects

were sampled at each location using pan trap sets consisting of one blue, one

yellow and one white bowl. Traps were made from plastic picnic bowls painted

with UV-bright fluorescent blue or yellow paint or left white (the reflectance of

each colour was not quantified in this study). Data for July (pre-flowering) were

collected from 24 sets of pan traps in almond orchards and 10 sets of pan traps in

native vegetation in Wemen and Colignan. Data for August-September (during

flowering) were collected from 57 sets of pan traps in almond orchards and 43

sets of pan traps in native vegetation. During the whole sampling period, a total of

19 individual pans (4.7%), five blue, seven yellow and seven white, were lost

from wind, weather or animal disturbance.

Sampling locations were at least 200 metres apart and were chosen a

priori from satellite maps to ensure a representative sample of the entire area of

almond orchards and so that native vegetation sites were located proximate to an

almond orchard. At all sites, sets of traps were placed on the ground, as close as

possible to a flowering tree, and filled with 200 mL of water and a drop of

dishwashing detergent to break surface tension. Very few studies advise at what

height pan traps were placed, and there appears to be no consensus on whether

traps should be mounted on poles (e.g. Tuell & Isaacs 2009) or placed on the

ground itself (e.g. Abrahamczyk et al. 2010). In this study, we placed pan traps on

the ground as the broader scope of the research was to investigate interactions

between potential pollinators and ground cover in almond orchards. Traps were

set out early in the morning (8—10 am) prior to maximum invertebrate activity,

and collected late in the afternoon just before nightfall (3.30—6 pm). Traps were

collected in the order they were placed to ensure that all traps were available to

insects for a similar time.

After collection, insects were removed from the pan traps and placed into

containers filled with 70% ethanol, before being transported back to the

laboratory. Insects were subsequently counted and identified to taxonomic order

and morphospecies, and stored in ethanol in refrigeration.

Data analysis

Data from both sampling periods were pooled and analyses were conducted using

PAST 2.14 (Hammer et al. 2001). The data set is ecological count data, with mild

168

heteroscedasticity (Brown-Forsythe test, P = 0.052) and a positively-skewed

distribution. Hence, we used non-parametric analysis on the untransformed data,

as evidence is mounting that it is preferable not to transform these type of data

(McArdle & Anderson 2004; O’Hara & Kotze 2010). Analyses were carried out

on combined counts of Hymenoptera and Diptera (Total Insects) and on Total

Diptera, Total Hymenoptera and subsets of Hymenoptera (European Honeybees,

Native Bees, and Total Native Hymenoptera).

Trap colour and insect catches

Friedman tests were performed to determine if a particular pan trap colour caught

more individuals across all habitat contexts for the following groups: Total

Insects, Total Hymenoptera, Total Diptera, Honeybees, Native Bees and Total

Native Hymenoptera. Post hoc multiple comparisons were conducted using

Wilcoxon signed rank tests for each trap colour pair (blue-yellow; yellow-white;

white-blue) to determine which colour caught the most individuals within each

pollinator insect group.

Habitat context and trap colour preference

The influence of habitat context on trap captures was analysed using two test

groups, Total Insects and Total Native Hymenoptera. Diptera were not included as

a separate group as 88% of our total captures were dipterans and initial analyses

on Total Diptera produced highly-similar results to analyses on Total Insects.

Native bees were combined with wasps to form the Total Native Hymenoptera

group, owing to the high number of zeroes in the Native Bee data set. In addition,

habitat context analysis of the Native Bee data set produced nearly identical

results to analysis of the Native Hymenoptera data set.

Analyses were conducted on the proportion of insects caught in a

particular trap colour across habitat contexts (native vegetation, MONO or

GRASS orchards). Proportions for each test group were calculated by dividing the

number of captures in a particular trap colour in a given habitat context by the

total number of trap captures (across all colours) in that habitat. This accounts for

the possibility that some habitat contexts may inherently support higher numbers

of insects than other habitats. Chi-squared contingency table analyses were

conducted to determine an overall effect and post-hoc pairwise comparisons were

169

calculated using the Marascuilo procedure (Marascuilo 1966) to determine

significant differences between each habitat pair combination (Native-MONO,

Native-GRASS, MONO-GRASS) for a given trap colour. For this procedure, the

rejection of H0 (that habitat has no influence on trap colour captures) occurs when

U'0 > Χ22,0.05 = 5.991.

RESULTS

Trap colour and insect catches

A total of 4060 Diptera and 436 Hymenoptera were captured in the study area.

Yellow bowls caught the most insects, while blue and white attracted relatively

equal numbers (Table A1.1). Results of Friedman tests revealed a statistically

significant effect of trap colour on the total number of insects caught as well as on

all individually-tested pollinator groups. Post-hoc pairwise comparisons showed

that yellow traps caught significantly more insects than blue or white (Table

A1.1). European honeybees were the only group that exhibited a significant

preference for blue traps over white (P = 0.035).

Habitat context and trap colour preference

Total Insects

Yellow traps caught the most pollinator insects in all habitats (Figure A1.2),

although habitat context significantly influenced the proportion of insects caught

in each trap colour (Table A1.2). Post-hoc comparisons showed that blue traps

were significantly more attractive in Native habitat than in either almond orchard

(Marascuilo procedure, Native vs MONO U'0 = 8.425, df = 2, P = 0.015; Native

vs GRASS: U'0 = 23.184, df = 2, P < 0.001). Conversely, white traps were

signficantly less attractive in Native habitat than in either orchard type (Native vs

MONO: U'0 = 6.517, df = 2, P = 0.038; Native vs GRASS: U'0 = 7.631, df = 2, P =

0.022) (Table A1.2).

170

Table A1.1: Pollinator insect abundance from pan trap collection in three habitat contexts (ncolour = 134) and results from Friedman tests (χ2) and Wilcoxon pairwise comparisons (W) to test for differences between number of individuals caught in each trap colour.

Pollinator Insect Group

Trap Colour χ2 W

Blue Yellow White Yellow/Blue Yellow/White White/Blue

Honeybee 64 88 23 15.351** 529* 690** 238.5*

Native Bee 10 57 8 19.043** 337* 268* 33

Other Native Hymenoptera 54 64 68 ~ ~ ~ ~

Total Native Hymenoptera 64 121 76 7.112* 1505* 1297* 699.5

Total Hymenoptera 128 209 99 11.352** 2460* 2425** 1073

Total Diptera 1150 1752 1158 16.112** 4928** 5176** 3735

Total Insects Mean (SE) Median

Interquartiles

1278

9.54 (1.3) 4

2, 9

1961

14.63 (2.22) 7

2, 15

1257

9.19 (1.26) 5

2, 9

24.609** 5388** 5050** 3588

Mean values indicate mean number of insects caught per trap of each colour. * P < 0.05 ** P < 0.001 ~ indicates that the analysis was not conducted on the associated variable.

171

Table A1.2: Results of Χ2 contingency table analysis and post-hoc pairwise comparisons (Marascuilo procedure) to determine whether habitat context influences insect captures in each coloured trap. Total Insect and Native Hymenoptera abundance in each trap colour is expressed as a proportion of the total individuals from that group caught in each habitat context.

TOTAL INSECTS NATIVE HYMENOPTERA

Χ24,0.05 27.491, P < 0.0001 15.609, P = 0.004

Blue Yellow White Blue Yellow White

Native 0.35 0.41 0.24 0.36 0.31 0.33

Native-MONO Χ22,0.05

8.4247,

P = 0.02

0.1439,

P = 0.93

6.5171,

P = 0.04

1.0842,

P = 0.58

2.856,

P = 0.24

0.6614,

P = 0.72

MONO 0.28 0.42 0.30 0.26 0.48 0.26

MONO-GRASS Χ22,0.05

0.2629,

P = 0.88

1.5233,

P = 0.47

0.6932,

P = 0.71

0.9812,

P = 0.61

0.5578,

P = 0.76

0.0133,

P = 0.99

GRASS 0.27 0.45 0.28 0.17 0.56 0.27

GRASS-Native Χ22,0.05 23.184, P<0.0001

4.6414,

P = 0.01

7.6308,

P = 0.02

9.3455,

P = 0.01

14.7964,

P = 0.0006

1.1159,

P = 0.57

172

Figure A1.2 Total pollinator insects caught in each coloured pan trap expressed as a proportion of total individuals caught in each habitat context (Native, n = 53; MONO, n = 42; GRASS, n = 39).

Native Hymenoptera

Yellow traps caught more native hymenopteran insects in GRASS and MONO

orchards than the other two colours, but native habitat had very little influence

over the attractiveness of different trap colours (Figure A1.3). Post-hoc

comparisons of each trap colour showed that blue traps were significantly more

attractive in Native habitats than in GRASS orchards (U'0 = 9.346, df = 2, P =

0.009). Conversely, yellow traps were significantly more attractive in GRASS

orchards than in Native habitats (U'0 = 14.79, df = 2, P = 0.0006).

Figure A1.3 Total Native Hymenoptera caught in each coloured pan trap expressed as a proportion of total individuals caught in each habitat context (Native, n = 53; MONO, n = 42; GRASS, n = 39).

173

DISCUSSION

Overall, we caught the most insect individuals (for all groups) in yellow traps.

Catches in blue and white traps were mostly similar across all insect groups,

except for European honeybees, which were significantly more abundant in blue

traps than in white. Habitat context significantly influenced pan trap catches. For

all insect groups combined, blue traps caught more individuals in native habitats

than in almond orchards, while white traps were more attractive within almond

orchards. For native hymenopterans, yellow traps caught more individuals in

GRASS orchards compared to native habitats, while blue traps were significantly

more attractive in native habitats.

Trap Colour

Pan trap surveys for pollinator insects highlight the wide variation in the

attractiveness of particular colours to different species or groups (e.g. Laubertie et

al. 2006; Cane et al. 2000; Gollan et al. 2011). Although most pan trap surveys

have focussed on bees, the few studies that include other pollinator insects (e.g.

flies, beetles, thrips etc.) have shown that one colour cannot be considered more

attractive than others when targeting a wide range of taxa in different ecological

contexts (Kirk 1984; Bowie et al. 1999; Vrdoljak & Samways 2012).

UV blue is currently considered to be the most suitable colour to attract

bees, perhaps because it has been noted previously that certain bee species in

some habitats prefer to visit blue or violet flowers (Kevan 1983; Weiss 2001).

Also, European honeybees have shown a preference for blue visual cues in some

laboratory experiments (e.g. Horridge 2007) and some pan trap studies from

North America have collected many more bee individuals in blue bowls than in

other coloured bowls (Cane et al. 2000; Campbell & Hanula 2007). Blue vane

traps (plastic jars with UV blue wing-like ‘vanes’ protruding from the mouth)

have also had some success attracting wild bee species in North America (Stephen

& Rao 2007; McKenney et al. 2009; Porter 2010) and have been recommended

for bee collecting by Stephen and Rao (2005). Only one published Australian

study has used blue vane traps to successfully sample native bees in an

agricultural landscape in southern New South Wales (Lentini et al.

2012).However, this study did not use any other coloured traps or insect sampling

methods, so it is still uncertain how effective blue vane traps are in Australian

174

ecosystems, relative to other trap types or colours. Other pan trapping surveys

have collected more bees in white bowls than blue (Grundel et al. 2011) or equal

numbers in blue and yellow bowls (Droege 2006) suggesting that bees do not

show a consistent preference for trap colour in all habitat contexts.

We collected a total of 682 more individuals in yellow traps than in blue

traps and both European honeybees and native bees were caught more frequently

in yellow traps, despite recommendations that the colour blue is more attractive to

bees (e.g. Cane et al. 2000; Stephen & Rao 2005). Nevertheless, in our sampled

habitats, honeybees were attracted to blue traps significantly more than native bee

species – 37% of honeybees were caught in blue traps compared to only 13% of

native bees. When compared with other Australian pan trap studies, our results

support the hypothesis that trap colour preference among pollinator insects,

including bees, is habitat-context specific. Gollan et al.’s (2011) pan trap survey

in New South Wales, which did not include blue bowls, caught more European

honeybees in white traps compared to yellow traps. In an unpublished

contemporaneous pan trap study of wet tropical rainforest and pasture systems in

north Queensland, the majority of captured bees were caught in blue pan traps;

however, the strength of colour preference varied among bee species, and across

seasons (Smith T, 2012, pers. comm. 25 April). In another Australian study,

which used five colours of sticky traps rather than bowls, Pickering and Stock

(2003) sampled alpine pollinators in the Mt Kosciuszko area and collected the

most insects in white traps. Hence, it is clear that variables other than trap colour

per se affect which colour, or combination of colours, is most attractive to target

insect groups in different habitat contexts.

Habitat context

The published literature on pan trap studies covers a variety of habitat contexts,

but a consistent preference for any one colour has not been determined. Results

have shown either no difference in catch abundances across bowl colours (e.g.

Toler et al. 2005; Wilson et al. 2008) or significant differences in catch

abundances (e.g. Campbell & Hanula 2007; Abrahamczyk et al. 2010). These

studies rarely correlate the pan trap data with ecological variables such as habitat

context, floral resource, vegetation structure or canopy cover (but see Missa et al.

175

2009; Abrahamczyk et al. 2010), so it is still unclear if the wide variety of pan

trap results is purely coincidental, or the result of an ecological interaction.

The habitat contexts sampled in our study differ in a variety of ways.

Almond plantations are managed as either bare-ground monocultures or as

orchards with continuous weedy/grassy ground cover. Almond orchards also have

a very different vegetation structure to the native mallee habitats. Although

MONO and GRASS orchards differ in that the latter has extensive ground cover,

they have essentially the same dimensional plant structure – all almond trees

within a block grow at essentially the same height and there is no understorey

vegetation. In contrast, native mallee habitats are more open than almond orchards

and have multiple vegetation layers, including a sparsely vegetated ground layer

and shrubs and trees ranging from two to twelve metres in height (Wood 1929).

We found that native hymenopterans were caught in yellow traps more often than

in blue or white traps within both MONO and GRASS orchards. However, at

mallee vegetation sites, hymenopterans showed no significant preference for any

colour. This could be because yellow traps were more visible to flying insects in

the orchard habitats, as a result of the contrast between trap colour and the

predominant colours of the surrounding habitat (i.e. green and white) (see

Laubertie et al. 2006; Abrahamczyk et al. 2010).

Almond trees are not native to Australia and produce white flowers,

sometimes tinged with pale pink. Weedy species in the GRASS orchards

produced mostly white and yellow flowers, but some other colours were observed,

both within the orchard and in adjacent patches. Mallee sites were typical of

Murray Mallee vegetation (Wood 1929), with a mixture of Eucalyptus, Callitris,

Acacia, Senna, Casuarina or Allocasuarina species. From personal observation,

the majority of flower-bearing plant species in these habitats produced yellow or

white flowers at the time of trapping. Blue flowers were rarely observed at native

mallee sites, yet blue traps caught significantly more insects in these locations

than in either orchard type.

The effect of local floral resources on the results of pan trap surveys has

been discussed previously. There is evidence to support theories of co-evolution

or adapted mutualisms between flowering plants and pollinators (Faegri & van der

Pijl 1979; Kearns et al. 1998; Chittka & Thomson 2001; Pauw & Bond,2011),

including between Australian Hymenoptera and native flowers (Dyer et al. 2012).

176

Although most pollinator insects are considered generalist foragers (Waser et al.

1996), it is also likely that the composition of local pollinator assemblages has

evolved in response to ecosystem variables, such as habitat structure, plant

diversity or predominant floral resource colour (Kevan et al. 2001; Pickering &

Stock 2003; Wenninger & Inouye 2008; Bates et al. 2011). Therefore, such

localised assemblages may respond to coloured traps differently depending on

their habitat context.

Very few studies have considered the interaction between pan trap colours

and floral colour in the sampled habitat. Leong and Thorp (1999) sampled bee

fauna in a flowering-vegetation patch dominated by white flowers and found that

females of the host-specific bee species preferred white bowls matching the colour

of their host plant, but other captured species preferred yellow bowls. Cane et al.

(2000) collected more bees in blue rather than yellow bowls, even though yellow

was the predominant flower colour at their study site, and Toler et al. (2005)

found there was no effect of the ecosystem-dominant flower colour on pan trap

catches of bees. An Australian study using six colours of pan traps found that

local floral resource influenced pan trap catches of hoverflies - when traps were

placed near yellow canola flowers, all hoverflies were caught in yellow traps, but

when traps were placed near white nashi flowers, 20% of sampled hoverflies were

caught in white bowls (Bowie et al. 1999). Although we did not statistically

quantify floral resources in this study, we caught more total insects in white traps

within both orchard types (where the dominant flower colour was white)

compared to native sites. In contrast, hymenopterans showed a significant

preference for yellow traps in GRASS orchards compared to native sites,

indicating that the dominant flower colour in those orchards did not influence this

taxonomic group.

Previous work considers some wasp species to see and respond to colours

differently to bees (Faegri & van der Pijl 1979; Peitsch et al. 1992; Lunau &

Maier 1995; but see also Desouhant et al. 2010), and an insect’s functional role

may also affect its response to colour, with different responses even appearing

within the same taxonomic family (Kirk 1984). Most experimental research into

insect colour vision has been conducted under controlled, laboratory conditions

(e.g. Horridge 2007; Desouhant et al. 2010) using European honeybees or

bumblebees (Lunau & Maier 1995; Kevan et al. 2001), both of which are native to

177

the northern hemisphere. Although these studies put much effort into designing a

variety of response contexts, these conditions can never emulate actual

environmental conditions across different habitats.

Colour has multiple aspects, such as intensity, wavelength, contrast, or

spectral purity, which will cause insects to respond differently depending on the

behavioural context (e.g. foraging, initial flower detection etc.) (Lunau & Maier

1995) or the environmental characteristics of the habitat itself, and these values

must not be considered as isolated variables (Menzel & Shmida 1993; Kevan et

al. 2001). For example, light has been shown to affect plant pollination, as insects

are less likely to visit shaded flowers compared to sunlit flowers (Kilkenny &

Galloway 2008). Light itself is also affected by numerous atmospheric factors

such as time of day, cloud cover, latitude or even hemisphere (Kevan 1983).

Radiation at ground level was not measured in this study, but light levels varied

between ‘interior’ sites (in the centre of almond orchard blocks) and more open

sites (edges of orchards or native vegetation). We also conducted this study during

late winter at latitude 34° S, both factors that will affect sunlight radiation and

levels of ambient light. Hence, it is likely that these variables would have

influenced the attractiveness of each pan trap colour in this study, and further

comparison across other seasons would confirm how significant such an effect

may be.

Conclusions and recommendations

Our results highlight the importance of considering the habitat context of

individual pan trapping surveys. Insects respond to colour differently across

environmental contexts and within taxonomic groups. Previous research has

shown that trap placement can interact with habitat context when attempting to

attract target insects in pan trapping surveys (Missa et al. 2009; Abrahamczyk et

al. 2010; Droege et al. 2010) and there is still much information to be gained on

the influence of habitat context on trapping success.

No one trap colour is likely to be suitable for trapping pollinator insects

across all habitat contexts. Rather, the most efficient use of pan trapping should

consider relevant environmental variables (e.g. vegetation structure), include

clusters of multiple colours, and address the potential for complementary

collection methods (e.g. sweep netting). Nevertheless, pan traps have immense

178

potential as a cheap, efficient collection method, especially for community- or

school-based science programs, and more sampling of this kind will provide

valuable ecological knowledge of insect diversity in different habitats, particularly

in the southern hemisphere.

ACKNOWLEDGEMENTS

Thank you to Select Harvests and Manna Farms for allowing access to their almond orchards; James Abell and Katrina Lumb for field assistance; MM Mayfield, T Smith and two anonymous reviewers for valuable comments on the draft manuscript; and GM Gurr for helpful guidance and encouragement during planning of the project. The Institute for Land Water & Society at Charles Sturt University provided a postgraduate scholarship to undertake the research project.

REFERENCES

Abrahamczyk, S., Steudel, B., Kessler, M. (2010) Sampling Hymenoptera along a precipitation gradient in tropical forests: the effectiveness of different coloured pan traps. Entomologia Experimentalis et Applicata 137:262-268.

Bates, A.J., Sadler, J.P., Fairbrass, A.J., Falk, S.J., Hale, J.D., Matthews, T.J. (2011) Changing bee and hoverfly pollinator assemblages along an urban-rural gradient. PLoS ONE 6:e23459.

Baum, K.A., Wallen, K.E. (2011) Potential bias in pan trapping as a function of floral abundance. Journal of the Kansas Entomological Society 84:155-159.

Boiteau, G. (1983) The arthropod community of potato fields in New Brunswick, 1979-1981. The Canadian Entomologist 115:847-853.

Bowie, M.H., Gurr, G.M., Hossain, Z., Baggen, L.R., Frampton, C.M. (1999) Effects of distance from field edge on aphidophagous insects in a wheat crop and observations on trap design and placement. International Journal of Pest Management 45:69-73.

Brunet, J. (2009) Pollinators of the Rocky Mountain columbine: temporal variation, functional groups and associations with floral traits. Annals of Botany103:1567-1578.

Bureau of Meteorology (2011) Climate Data Online. Available from URL: http://www.bom.gov.au/climate/data/.

Campbell, J.W., Hanula, J.L. (2007) Efficiency of Malaise traps and coloured pan traps for collecting flower visiting insects from three forested ecosystems. Journal of Insect Conservation 11:399-408.

179

Cane, J.H., Minckley, R.L., Kervin, L.T. (2000) Sampling bees (Hymenoptera: Apiformes) for pollinator community studies: pitfalls of pan-trapping. Journal of the Kansas Entomological Society 73:225-231.

Chittka, L., Thomson, J.D. (eds.) (2001) Cognitive Ecology of Pollination. Cambridge University Press,Cambridge.

Desouhant, E., Navel, S., Foubert, E., Fischbein, D., Théry, M., Bernstein, C. (2010) What matters in the associative learning of visual cues in foraging parasitoid wasps: colour or brightness? Animal Cognition 13:535-543.

Droege, S. (2006) Impact of colour and size of bowl trap on numbers of bees captured. Available from URL: http://online.sfsu.edu/~beeplot/pdfs/color%20and%20size.pdf.

Droege, S., Tepedino, V.J., Lebuhn, G. et al. (2010) Spatial patterns of bee captures in North American bowl trapping surveys. Insect Conservation & Diversity 3:15-23.

Dyer, A.G., Boyd-Gerny, S., McLoughlin, S., Rosa, M.G.P., Simonov, V., Wong, B.B.M. (2012) Parallel evolution of angiosperm colour signals: common evolutionary pressures linked to hymenopteran vision. Proceedings of the Royal Society B, doi: 10.1098/rspb.2012.0827.

Evans, D.A., Medler, J.T. (1967) Flight activity of corn leaf aphid in Wisconsin as determined by yellow pan trap collections. Journal of Economic Entomology60:1088.

Faegri, K., van der Pijl, L. (1979) The Principles of Pollination Ecology. Pergamon Press Ltd., Oxford.

Gollan, J.R., Ashcroft, M.B., Batley, M. (2011) Comparison of yellow and white pan traps in surveys of bee fauna in New South Wales, Australia (Hymenoptera: Apoidea: Anthophila). Australian Journal of Entomology 50:174-178.

Grundel, R., Frohnapple, K.J., Jean, R.P., Pavlovic, N.B. (2011) Effectiveness of bowl trapping and netting for inventory of a bee community. Environmental Entomology 40:374-380.

Hammer, Ø., Harper, D.A.T., Ryan, P.D. (2001) PAST: Paleontological statistics software package for education and data analysis. Paeleontologia Electronica4(1):9pp. Available from URL: http://palaeo-electronica.org/2001_1/past/issue1_01.htm.

Horridge, A. (2007) The preferences of the honeybee (Apis mellifera) for different visual cues during the learning process. Journal of Insect Physiology 53:877-889.

Jauker, F., Diekötter, T., Schwarzbach, F., Wolters, V. (2009) Pollinator dispersal in an agricultural matrix: opposing responses of wild bees and hoverflies to

180

landscape structure and distance from main habitat. Landscape Ecology 24:547-555.

Kearns, C.A., Inouye, D.W., Waser, N.M. (1998) Endangered Mutualisms: the conservation of plant-pollinator interactions. Annual Review of Ecology & Systematics 29:83-112.

Kevan, P.G. (1983) Floral colours through the insect eye: what they are and what they mean. In: Handbook of Experimental Pollination Biology (eds CE Jones & RJ Little) pp. 3-30. Scientific and Academic Editions/Van Nostrand Reinhold Company Inc., New York

Kevan, P.G., Baker, H.G. (1983) Insects as flower visitors and pollinators. Annual Review of Entomology 28:407-453.

Kevan, P.G., Chittka, L., Dyer, A.G. (2001) Limits to the salience of ultraviolet: lessons from colour vision in bees and birds. Journal of Experimental Biology204:2571-2580.

Kilkenny, F.F., Galloway, L.F. (2008) Reproductive success in varying light environments: direct and indirect effects of light on plants and pollinators. Oecologia 155:247-255.

Kirk, W.D.J. (1984) Ecologically selective coloured traps. Ecological Entomology9:35-41.

Klein, A-M., Vaissière, B.E., Cane, J.H., Steffan-Dewenter, I., Cunningham, S.A., Kremen, C., Tscharntke, T. (2007) Importance of pollinators in changing landscapes for world crops. Proceedings of the Royal Society B 274:303-313.

Larson, B.M.H., Kevan, P.G., Inouye, D.W. (2001) Flies and flowers: taxonomic diversity of anthophiles and pollinators. The Canadian Entomologist 133:439-465.

Laubertie, E.A., Wratten, S.D., Sedcole, J.R. (2006) The role of odour and visual cues in the pan-trap catching of hoverflies (Diptera: Syrphidae). Annals of Applied Biology 148:173-178.

Lentini, P.E., Martin, T.G., Gibbons, P., Fischer, J., Cunningham, S.A. (2012) Supporting wild pollinators in a temperate agricultural landscape: Maintaining mosaics of natural features and production. Biological Conservation 149:84-92.

Leong, J.M., Thorp, R.W. (1999) Colour-coded sampling: the pan trap colour preferences of oligolectic and nonoligolectic bees associated with a vernal pool plant. Ecological Entomology 24:329-335.

Lunau, K., Maier, E.J. (1995) Innate colour preferences of flower visitors. Journal of Comparative Physiology A 177:1-19.

Marascuilo, L.A. (1966) Large-sample multiple comparisons. Psychological Bulletin 65:280-290.

181

McArdle, B.H., Anderson, M.J. (2004) Variance heterogeneity, transformations, and models of species abundance: a cautionary tale. Canadian Journal of Fisheries and Aquatic Sciences 61:1294-1302.

McKenney, M.A., Rao, S., Stephen, W.P. (2009) Pollination in Cranberries on the South Oregon Coast: Honeybees. South Coast Grower News, vol. 3, no. 1, Spring 2009, pp 3-5. Available from URL: http://ir.library.oregonstate.edu/xmlui/bitstream/handle/1957/16200/McKenneyGrowerNewsSpring2009.pdf?sequence=1.

Menzel, R., Shmida, A. (1993) The ecology of flower colours and the natural colour vision of insect pollinators: the Israeli flora as a study case. Biological Reviews 68:81-120.

Missa, O., Basset, Y., Alonso, A., et al. (2009) Monitoring arthropods in a tropical landscape: relative effects of sampling methods and habitat types on trap catches. Journal of Insect Conservation 13:103-118.

Natural Resources & Environment. (n.d.) Victorian Mallee Bioregions. Victoria’s Biodiversity, Department of Natural Resources & Environment, Victoria. URL: http://www.nre.vic.gov.au/plntanml/biodiversity/directions/mallee.htm.

O’Hara, R.B., Kotze, D.J. (2010) Do not log-transform count data. Methods in Ecology & Evolution 1:118-122.

Pauw, A., Bond, W.J. (2011) Mutualisms matter: pollination rate limits the distribution of oil-secreting orchids. Oikos 120:1531-1538.

Peitsch, D., Fietz, A., Hertel, H., de Souza, J., Ventura, D.F., Menzel, R. (1992) The spectral input systems of hymenopteran insects and their receptor-based colour vision. Journal of Computational Physiology A 170:23-40.

Pickering, C.M., Stock, M. (2003) Insect colour preference compared to flower colours in the Australian Alps. Nordic Journal of Botany 23:217-223.

Porter, D.A. (2010) An investigation of wild bee diversity and abundance in plots managed by The Nature Conservancy in south-central Nebraska and of beneficial arthropods associated with native Nebraska flora. Thesis, Master of Science, University of Nebraska.

Rader, R., Edwards, W., Westcott, D.A., Cunningham, S.A., Howlett, B.G. (2011) Pollen transport differs among bees and flies in a human-modified landscape. Diversity & Distributions 17:519-529.

Roulston, T.H., Smith, S.A., Brewster, A.L. (2007) A comparison of pan trap and intensive net sampling techniques for documenting a bee (Hymenoptera: Apiformes) fauna. Journal of the Kansas Entomological Society 80:179-181.

182

Stephen, W.P., Rao, S. (2005) Unscented colour traps for non-Apis bees (Hymenoptera: Apiformes). Journal of the Kansas Entomological Society 78:373-380.

Stephen, W.P., Rao, S. (2007) Sampling native bees in proximity to a highly competitive food resource (Hymenoptera: Apiformes). Journal of the Kansas Entomological Society 80:369-376.

Toler, T.R., Evans, E.W., Tepedino, V.J. (2005) Pan-trapping for bees (Hymenoptera: Apiformes) in Utah’s West Desert: the importance of colour diversity. Pan-Pacific Entomologist 81:103-113.

Tuell, J.K., Isaacs, R. (2009) Elevated pan traps to monitor bees in flowering crop canopies. Entomologia Experimentalis et Applicata, 131:93-98.

Vrdoljak, S.M., Samways, M.J. (2012) Optimising coloured pan traps to survey flower visiting insects. Journal of Insect Conservation 16:345-354.

Waser, N.M., Chittka, L., Price, M.V., Williams, N.M., Ollerton, J. (1996) Generalisation in pollination systems, and why it matters. Ecology 77:1043-1060.

Weiss, M.R. (2001) Vision and learning in some neglected pollinators: beetles, flies, moths, and butterflies. In: Cognitive Ecology of Pollination (eds L Chittka & JD Thomson) pp. 171-190. Cambridge University Press, Cambridge.

Wenninger, E.J., Inouye, R.S. (2008) Insect community response to plant diversity and productivity in a sagebrush-steppe ecosystem. Journal of Arid Environments 72:24-33.

Westphal, C., Bommarco, R., Carré, G., et al. (2008) Measuring bee diversity in different European habitats and biogeographical regions. Ecological Monographs78:653-671.

Wilson, J.S., Griswold, T., Messinger, O.J. (2008) Sampling bee communities (Hymenoptera: Apiformes) in a desert landscape: are pan traps sufficient? Journal of the Kansas Entomological Society 81:288-300.

Wood, J.G. (1929) Floristics and Ecology of the Mallee. Transactions of the Royal Society of South Australia 53:59-78.

183

Appendix 2: Taxonomic lists of flower visitor observations and trap samples

Table A2.1 Observations of insects visiting open flowers at almond and mallee sites, August-September 2010. Sites are grouped by vegetation type (GRASS = plant-diverse almond; MONO = monoculture almond; Native = mallee vegetation).

GRASS

(n = 27)

MONO

(n = 30)

Native

(n = 43)

Diptera (unidentified) 38 4 8

Diptera: Syrphidae 3 5

Hymenoptera: Formicidae 2

Hymenoptera (all wasps) 2 15

Lepidoptera (unidentified) 1 17

Site notes Honeyeaters and other small birds also occasional feeders on almond flowers.

Fly individual on almond flower at Wemen plantation seen with pollen grains attached to leg.

N.B. Data collection – random walks within 10 metres of each trap location, for a 10 minute period; European honeybees were observed at all vegetation types, but not recorded

184

Figure A2.1 Representative photographs of Diptera, Lepidoptera and Hymenoptera individuals observed visiting flowers in almond plantations (top row) and native vegetation (bottom row). See Figure 2.2 on page 13 for additional photographs of pollinators at native vegetation sites.

185

Table A2.2 Total abundance of European honeybees, native bees (species), wasps and flies (family groups) collected at almond (GRASS, MONO) and Native mallee vegetation sites, August-September 2010 (Chapters 3 & 4).

GRASS Native MONO

Nangiloc

(n =8)

Colignan B

(n =9)

Colignan C

(n =10)

NC (Colignan)

(n =13)

NV (Wemen)

(n =15)

NL (Liparoo)

(n = 15)

Wemen

(n =15)

Liparoo

(n =15)

HYMENOPTERA

Apis mellifera 10 23 12 3 10 61 11 45

Total native bees 5 7 49 10 15 3

Homalictus sphecoides 2

H. urbanus 2 2 4 2 1

Lasioglossum (Chilalictus) athrix 13

L. (C.) brunnesetum 2 2

186

L. (C.) callophyllae 5

L. (C.) erythrurum 1

L. (C.) globosum 22 3

L. (C.) helichrysi 3

L. (C.) victoriae 1

L. (C.) spp. 2 4 10 2

Leioproctus spp. 1 1 2 1

Lipotriches spp. 1

187

Total wasps 24 20 32 14 16 9 19 10

Bethylidae 1

Braconidae 10 4 3 11 1 7

Ceraphronidae 1 1

Chalcididae 1 2 1 1 1

Crabronidae 1 1 1

Diapriidae 6 12 12 2 1 5 6 4

Evaniidae 1 1

Ichneumonidae 2 1

188

Megaspilidae 1 1

Mymaridae 1 1

Perilampidae 1

Platygastridae 1

Proctotrupidae 7 1 3 1

Pteromalidae 2 1 2 1 1 3

Scelionidae 3 3 1 1 2

Sphecidae 3

Tiphiidae 1

Trichogrammatidae 1

189

DIPTERA (total) 1185 968 201 274 138 139 139 207

Asilidae 1 11 10 3

Bombyliidae 1 2

Calliphoridae 238 79 14 17 10 12 44 15

Ceratopogonidae 57 26 8 7 3 15 14

Chironomidae 6 35 7 4 9 1 11 6

Chloropidae 29 29 19 12 3 1 3 4

Cryptochetidae 1

Dolichopodidae 20 5 11 3 4 4

Drosophilidae 64 231 27 8 4 18 7 21

Ephydridae 2 16 6 2 2 1 4

Heleomyzidae 17 24 10 1 10 1 11 19

190

Hybotidae 1 1

Milichiidae 7 6

Muscidae 1 5 3 3 1 1 3

Phoridae 75 85 27 10 11 11 12 10

Pipunculidae 1

Platypezidae 3 2 33 5 10 6 2 1

Sarcophagidae 1 3

Sepsidae 1

Stratiomyidae 1 5 3

Syrphidae 1 3 2 8 1

Tachinidae 689 429 21 190 39 59 27 104

Tephritidae 1

191

Therevidae 2 1 1

Trichoceridae 2 1

192

Table A2.3 Total abundance of European honeybees, native bees (species), wasps and flies (family groups) collected at monoculture almond plantation and native mallee vegetation sites, July-September 2011 (Chapters 5 & 6).

July August September

Almond

(n = 50)

Mallee

(n = 18)

Almond

(n = 50)

Mallee

(n = 18)

Almond

(n = 50)

Mallee

(n = 18)

HYMENOPTERA

Apis mellifera 1 5 3 2

Total Native Bee 4 86 81

Lasioglossum (chilalictus) spp. 3 79 78

L (Ch.) erythrurum 1

Homalictus urbanus 1 6 3

Total Wasps (inc. Symphyta) 30 28 44 43 128 63

Bethylidae 1 1 8

193

Braconidae 3 1 3 2 9 1

Ceraphronidae 2 2 3 8 4

Chalcididae 2

Crabronidae 2 4

Cynipidae 4

Diapriidae 6 1 5 2 24 1

Eulophidae 1 13 4

Eurytomidae 2 1

Evaniidae 4

Figitidae 1 3 2

194

Gasteruptiidae 1

Ichneumonidae 1 1 3 10 8

Megaspilidae 3 2

Mymaridae 4 4 7 2

Pergidae 4 19

Perilampidae 1

Platygastridae 2 6 1

Pompilidae 17

Proctotrupidae 1 3

Pteromalidae 1 5 10 5 14 10

Scelionidae 11 8 8 3 15 4

195

Sphecidae 1 1

Tiphiidae 5 2 1 1

DIPTERA (total) 157 125 847 179 974 112

Asilidae 4 25 1 9 6 15

Calliphoridae 7 18 14 30 5

Ceratopogonidae 5

Chironomidae 30 7 52 6 65 2

Chloropidae 5 111 9 157 16

Dolichopodidae 7 3 204 7 363 17

Drosophilidae 12 2 64 13 68

Heleomyzidae 13 1 14 60 12

196

Milichiidae 2 2

Muscidae 4 2 10 11

Phoridae 33 3 105 18 69 1

Platypezidae 3 6 5 6 4

Sarcophagidae 8 2

Sepsidae 5 1

Syrphidae 1 2 3

Tachinidae 45 68 145 80 114 38

Tephritidae 1 2 1

Therevidae 1

Unidentified Acalyptrata 55 14

Unidentified Brachycera 54 11

197

Appendix 3: Full set of models for analyses of ground cover vegetation—pollinator relationships (Chapter 3)

Table A3.1 Parameter estimates with confidence intervals from negative binomial regression (abundance) and linear regression (richness) analysis testing the relationship between insect groups and ground cover variables. Akaike weights (wi) for each model indicate the probability that the given model is the best fit for the data. Insect response Model AICc Δ AICc wi

Fly Abundance Constant only (-0.11, -0.14 to -0.09)

D (0, 0 to 0)

PR (-0.005, -0.02 to 0.007)

GC (0, -0.001 to 0.001)

PR (-0.005, -0.02 to 0.008) + D (0, 0 to 0)

GC (0, -0.001 to 0.001) + D (0, 0 to 0)

PR (-0.003, -0.02 to 0.02) + D (0, 0 to 0) + PR*D (0, 0 to 0)

GC (0, -0.001 to 0.001) + D (0, 0 to 0) + GC*D (0, 0 to 0)

325.5

327.1

327.1

327.6

328.9

329.3

331.1

331.3

0

1.5

1.6

2.1

3.4

3.8

5.6

5.7

0.37

0.17

0.16

0.13

0.07

0.06

0.02

0.02

198

Richness GC (-0.003, -0.007 to 0.001)

Constant only (1.33, 1.18 to 1.49)

PR (-0.06, -0.15 to 0.05)

D (0, 0 to 0)

GC (-0.003, -0.007 to 0.002) + D (0, -0.001 to 0.001)

PR (-0.06, -0.15 to 0.05) + D (0, -0.001 to 0)

GC (-0.004, -0.01 to 0.004) + D (0, 0 to 0) + GC*D (0, 0 to 0)

PR (-0.09, -0.22 to 0.07) + D (0, -0.002 to 0.001) + PR*D (0, -0.001 to 0)

107.9

108.1

108.4

109.7

109.8

110.3

111.9

112.5

0

0.2

0.5

1.8

1.9

2.4

4.1

4.6

0.25

0.23

0.20

0.10

0.10

0.07

0.03

0.03

Wasp Abundance GC (0.006, 0.003 to 0.01)

GC (0.007, 0.003 to 0.01) + D (0.001, -0 to 0.002)

GC (0.01, 0 to 0.01) + D (0.001, -0.001 to 0.003) + GC*D (0, 0 to 0)

PR (0.12, 0.04 to 0.19) + D (0.001, 0 to 0.002)

201.9

202.3

204.5

206.5

0

0.4

2.6

4.6

0.43

0.34

0.12

0.04

199

PR (0.10, 0.03 to 0.17)

PR (0.06, -0.065 to 0.19) + D (0, -0.003 to 0.002) + PR*D (0, 0 to 0.001)

206.8

207.5

4.9

5.5

0.04

0.03

Richness GC (0.006, 0.001 to 0.01)

GC (0.006, 0.002 to 0.01) + D (0.001, 0 to 0.002)

PR (0.12, 0.008 to 0.22)

PR (0.13, 0.02 to 0.23) + D (0.001, 0 to 0.003)

GC (0.005, -0.003 to 0.01) + D (0.001, -0.002 to 0.003) + GC*D (0, 0 to 0)

PR (0.08, -0.09 to 0.02) + D (0, -0.002 to 0.003) + PR*D (0, -0.001 to

0.001)

Int (0.87, 0.68 to 1.1)

D (0.001, -0.001 to 0.002)

127.9

128.3

129.7

129.9

130.5

131.7

131.8

132.7

0

0.4

1.8

2.0

2.6

3.8

3.9

4.8

0.31

0.26

0.13

0.11

0.09

0.05

0.05

0.03

Native bee Abundance PR (1.18, 0.52 to 1.8) 88.0 0 0.60

200

PR (1.2, 0.49 to 1.8) + D (0, -0.006 to 0.005)

GC (0.05, 0.02 to 0.07)

PR (0.98, -0.02 to 2.0) + D (-0.004, -0.02 to 0.01) + PR*D (0.001, -0.004 to

0.007)

GC (0.04, 0.02 to 0.07) + D (-0.003, -0.008 to 0.003)

GC (0.04, -0.01 to 0.09) + D (-0.004, -0.02 to 0.01) + GC*D (0, 0 to 0)

90.2

92.0

92.4

93.6

95.9

2.2

4.0

4.4

5.6

7.9

0.20

0.08

0.07

0.04

0.01

201

Richness GC (0.007, 0.004 to 0.01)

GC (0.007, 0.005 to 0.01) + D (0, 0 to 0.001)

GC (0.005, 0 to 0.01) + D (0, -0.001 to 0.001) + GC*D (0, 0 to 0)

PR (0.14, 0.07 to 0.21)

PR (0.14, 0.07 to 0.21) + D (0, 0 to 0.001)

PR (0.08, -0.02 to 0.19) + D (0, -0.002 to 0.001) + PR*D (0, o to 0.001)

72.2

73.6

74.6

79.8

81.1

81.2

0

1.4

2.4

7.6

8.9

9.0

0.54

0.27

0.17

0.01

0.006

0.006

GC = percent ground cover; PR = plant richness; D = distance to natural or semi-natural vegetation. All models include the constant as a parameter.

202

Appendix 4: Location attributes and pollinator occurrence along ecotone gradients (Chapter 5)

Table A4.1 Physical and historical attributes of the studied ecotone patches. Name Type Patch Size (ha) Management History

Liparoo Commercial almond orchard 1002 orchard established 2001, broadacre conventional

monoculture (herbicide only)

Annuello Regenerated mallee woodland 35283.5 State flora and fauna reserve, established 1990

Wemen Commercial almond orchard 165 orchard established 1988, broadacre conventional

monoculture (herbicide only)

Collins Remnant mallee woodland 210 unmanaged mallee remnant, unknown history

203

Table A4.2 Monthly occurrence of Hymenoptera and Diptera families along the (a) Annuello—Liparoo and (b) Collins—Wemen ecotones. The

shaded portion of the table represents the mallee side of the boundary. J = July, A = August, S = September. Bold family names indicate the family was

unique to the specified location (i.e. only collected at either Annuello/Liparoo or Collins/Wemen).

Family (a) Distance along Annuello—Liparoo ecotone (m)

-300 -40 -10 10 40 70 100 300

HYMENOPTERA

Bethylidae A S S

Braconidae A A S A,S

Ceraphronidae A,S S J,A A S

Crabronidae S S

Cynipidae J J

Diapriidae J,A,S S S A,S J,S

204

Eulophidae S S S S

Figitidae A A

Halictidae S S S S S S S A,S

Ichneumonidae A,S A S S A

Megaspilidae A A

Mymaridae S S A S

Perilampidae S

Platygastridae J J,S S

Pompilidae S S

Proctotrupidae J S S

Pteromalidae A,S J,A A,S A,S A,S A,S S

Scelionidae J J,S J,A,S S J,A,S J,S A,S J,S

205

Sphecidae S

Tiphiidae J,A J,S S

DIPTERA

Asilidae S J S S S

Calliphoridae J,A,S A A,S A,S S

Chironomidae J,A J S A,S J,A,S J,A,S

Chloropidae S A,S A,S A,S A,S A,S A,S J,A,S

Dolichopodidae A J,A,S A,S J,S J,A,S J,A,S A,S A,S

Drosophilidae A A J,A,S J,A A A J,A,S

Heleomyzidae S S J,A,S J,A,S J,S J,S J,A,S

Milichiidae A A

Muscidae J A A S A,S J,A

206

Phoridae A A J,A J,A,S J,A J,A,S J,A,S J,A,S

Platypezidae J J,A A J J J

Sarcophagidae S S

Sepsidae A A A

Syrphidae S S A,S

Tachinidae A,S J,A,S J,A,S J,A,S J,A,S J,A,S J,A,S J,A,S

Tephritidae S

un. Acalyp A

un. LBrach S

un. Acalyp = unidentified Acalyptrata; un. LBrach = unidentified Lower Brachycera

207

Family

(b) Distance along Collins—Wemen ecotone (m)

-300 -40 -10 10 40 70 100 300

HYMENOPTERA

Bethylidae J S S

Braconidae S J J J,A,S J,S A

Ceraphronidae A S J,S

Chalcididae S S

Crabronidae S S S

Cynipidae J

Diapriidae A S J,A A,S S J,S S

Eulophidae S J S S S

208

Eurytomidae S S S

Evaniidae A A

Figitidae S A J,S

Gasteruptiidae S

Halictidae S A,S S S A,S A,S S S

Ichneumonidae S J,S S S S S

Megaspilidae S A,S

Mymaridae J J S A S A,S A S

Pergidae A A A A

Platygastridae S S S

Pompilidae S S

Proctotrupidae J S S

209

Pteromalidae J,S J,A J,S A J,A,S A,S S

Scelionidae S J J,S J,A S A A,S

Sphecidae S

DIPTERA

Asilidae J,A,S J,A J,S A,S

Calliphoridae J,A,S J,A J,A,S A,S A,S J,A,S A,S J,A,S

Ceratopogonidae S S S

Chironomidae J,S A,S A J,A,S J,A,S J,A,S J,A,S J,A,S

Chloropidae A S J,A,S A,S A,S A,S A,S

Dolichopodidae J,A,S J,A,S J,A,S J,A,S A,S A,S A,S

Drosophilidae J J,A,S A,S J,A,S J,A,S S

Heleomyzidae S J,S S A,S S S

210

Milichiidae S S

Muscidae S J, S J J,S

Phoridae A A,S A J,A,S J,A,S J,A,S J,A,S J,A,S

Platypezidae A S A A,S S

Sarcophagidae S S

Sepsidae A

Syrphidae J,A

Tachinidae J,A,S J,A,S J,A,S J,A,S J,A,S J,A,S J,A,S J,A,S

Tephritidae J,A A

Therevidae S

un. Acalyp A A A A A A A

un. LBrach A,S S A A A

211

Appendix 5: Physical attributes and full candidate sets of models for analysis of habitat structure relationships (Chapter 6)

Table A5.1 Location and physical attributes of the sampled monoculture almond plantations. Liparoo Wemen

Location -34.818° 142.576° -34.751° 142.679°

Year established 2001 1988

Area (ha) 1,002 165

Area of adjacent mallee patch (ha) 35,000 – Annuello Flora and Fauna Reserve 210 – unnamed patch

Average height of almond trees 5.3 m SE 0.17 m 6.6 m SE 0.13 m

Row width 5-6 m 5-6 m

Tree spacing 5 m 5 m

212

Table A5.2 The candidate set of models (Δi AICc ≤ 10) explaining relationships between habitat attributes and monthly wild pollinator abundance (A)

or richness (R) across all almond plantation and mallee woodlands sites (n = 68).

Response Month Model Δi AICc L(gi) wi

Fly (A) July Veg (0.47; 1.2) + GC (-15.65; 9.8) + Veg*GC (49.04; 17.1)

Veg (1.64; 2.2) + GC (-13.81; 9.7) + LL (16.78; 9.6) + Veg*GC (49.19; 18.9) + Veg*LL (-19.6; 14.0)

GC (15.5; 8.4) + LL (11.23; 5.6)

Veg (3.72; 1.2)

GC (26.49; 9.3)

Veg (4.26; 2.4) + LL (16.68; 9.8) + Veg*LL (-16.11; 14.1)

LL (16.21; 5.5)

0

1.36

1.87

3.11

3.88

4.36

5.15

1

0.51

0.39

0.21

0.14

0.11

0.08

0.41

0.21

0.16

0.09

0.06

0.05

0.03

213

August Veg (-6.78; 2.2)

LL (-28.99; 8.2)

Veg (0.63; 4.3) + LL (72.76; 50.1) + Veg*LL (-94.53; 51.8)

GC (-17.36; 16.5) + LL (-25.65; 9.2)

Constant only (15.27; 1.2)

GC (-26.07; 12.9)

Veg (-5.72; 2.8) + GC (30.15; 61.0) + Veg*GC (-38.03; 62.9)

Veg (2.67; 5.2) + GC (41.08; 59.6) + LL (74.44; 47.8) + Veg*GC (-54.77; 62.78) + Veg*LL (-97.67; 49.8)

0

0.31

0.46

1.81

3.61

3.73

4.03

4.39

1

0.86

0.80

0.41

0.17

0.16

0.13

0.11

0.28

0.24

0.22

0.11

0.05

0.04

0.04

0.03

September Veg (-0.76; 0.2) + GC (5.56; 2.6) + Veg*GC (-7.37; 3.0)

Veg (-1.32; 0.3) + LL (3.07; 1.9) + Veg*LL (-1.06; 2.2)

Veg (-0.99; 0.2)

Veg (-1.54; 0.3) + CC (0.12; 0.5) + GC (1.39; 1.3) + LL (2.63; 0.9)

Veg (-0.51; 0.5) + CC (0.29; 0.5) + Veg*CC (-1.52; 1.4)

0

0.38

1.34

1.84

4.71

1

0.83

0.51

0.40

0.10

0.35

0.29

0.18

0.14

0.03

214

(R) July LL (1.27; 0.7)

Constant only (1.50; 0.1)

GC (0.69; 1.1) + LL (1.18; 0.7)

Veg (0.21; 0.2)

GC (1.10; 1.1)

Veg (0.01; 0.2) + GC (-4.67; 3.2) + Veg*GC (6.23; 3.6)

Veg (0.02; 0.3) + LL (1.85; 2.2) + Veg*LL (-0.62; 2.5)

Veg (-0.46; 0.4) + GC (-4.45; 3.2) + LL (1.56; 2.2) + Veg*GC (7.08; 3.5) + Veg*LL (0.48; 2.5)

0

0.96

1.26

1.72

2.22

3.21

4.52

4.86

1

0.62

0.53

0.42

0.33

0.20

0.10

0.09

0.30

0.19

0.16

0.13

0.10

0.06

0.03

0.03

215

August Constant only (2.12; 0.1)

Veg (-0.13; 0.2) + GC (-7.84; 3.2) + Veg*GC (7.96; 3.5)

GC (-1.01; 1.1)

LL (-0.34; 0.7)

Veg (-0.04; 0.2)

Veg (0.36; 0.4) + GC (-7.51; 3.1) + LL (2.37; 2.1) + Veg*GC (6.84; 3.5) + Veg*LL (-3.88; 2.5)

GC (-0.94; 1.2) + LL (-0.21; 0.7)

Veg (0.35; 0.3) + LL (2.86; 2.2) + Veg*LL (-4.16; 2.5)

0

0.88

1.39

1.97

2.13

3.01

3.57

3.84

1

0.65

0.50

0.37

0.34

0.22

0.17

0.15

0.29

0.19

0.15

0.11

0.10

0.07

0.05

0.04

216

September Veg (-0.8; 0.3) + CC (-1.05; 0.5) + GC (0.19; 1.2) + LL (2.06; 0.9)

Veg (-0.63; 0.3) + LL (0.12; 1.9) + Veg*LL (2.47; 2.2)

Constant only (1.92; 0.1)

CC (-0.38; 0.3)

LL (0.73; 0.6)

GC (-1.60; 1.1) + CC (-0.59; 0.4)

Veg (-0.50; 0.4) + CC (-1.12; 0.5) + Veg*CC (0.19; 1.2)

GC (-0.91; 1.0)

GC (-1.22; 1.0) + LL (0.89; 0.7)

Veg (-0.09; 0.2)

LL (0.47; 0.8) + CC (-0.25; 0.4)

CC (-0.45; 0.4) + GC (-1.61; 1.1) + LL (0.49; 0.7)

Veg (0.07; 0.2) + GC (1.99; 2.9) + Veg*GC (-3.45; 3.2)

0

1.54

1.58

2.49

2.49

2.56

2.86

2.96

3.33

3.43

4.36

4.46

6.40

1

0.46

0.45

0.29

0.29

0.28

0.24

0.23

0.19

0.18

0.11

0.11

0.04

0.26

0.12

0.12

0.07

0.07

0.07

0.06

0.06

0.05

0.05

0.03

0.03

0.01

217

Wasp (A) July LL (4.45; 2.0)

Veg (0.96; 0.4)

GC (2.57; 2.7) + LL (3.9; 1.9)

Constant only (0.85; 0.2)

GC (4.61; 3.4)

Veg (0.43; 0.7) + LL (2.39; 4.2) + Veg*LL (0.58; 5.5)

Veg (1.07; 0.7) + GC (2.81; 4.9) + Veg*GC (-3.65; 6.9)

Veg (0.15; 1.0) + GC (2.88; 4.9) + LL (2.36; 3.9) + Veg*GC (-1.25; 6.6) + Veg*LL (1.36; 5.5)

0

0.04

1.43

2.18

2.65

3.89

4.23

8.30

1

0.98

0.49

0.34

0.27

0.14

0.12

0.02

0.30

0.29

0.15

0.10

0.08

0.04

0.04

0.01

218

August GC (1.95; 0.7)

Veg (0.71; 0.3) + LL (0.26; 2.9) + Veg*LL (-1.89; 3.1)

Veg (0.41; 0.2)

GC (1.95; 0.7) + LL (-0.01; 0.7)

Veg (0.30; 0.2) + GC (1.91; 3.5) + Veg*GC (-0.71; 3.5)

Veg (0.67; 0.3) + GC (2.09; 3.6) + LL (0.44; 3.0) + Veg*GC (-1.64; 3.7) + Veg*LL (-1.91; 3.2)

Constant only (-0.47; 0.1)

LL (0.11; 0.6)

0

0.15

0.37

2.19

2.23

4.21

6.18

8.28

1

0.93

0.83

0.33

0.33

0.12

0.05

0.02

0.28

0.26

0.23

0.09

0.09

0.03

0.01

0.004

219

September Constant only (-0.49; 0.1)

Veg (0.38; 0.2) + LL (0.54; 1.5) + Veg*LL (-1.95; 1.7)

Veg (0.12; 0.1)

GC (0.66; 0.6)

CC (-0.19; 0.2)

LL (-0.27; 0.5)

LL (-0.59; 0.5) + CC (-0.34; 0.3)

GC (0.73; 0.5) + LL (-0.38; 0.5)

GC (0.55; 0.6) + CC (-0.11; 0.2)

Veg (0.04; 0.1) + GC (-2.43; 2.9) + Veg*GC (3.05; 3.1)

CC (-0.27; 0.3) + GC (0.51; 0.6) + LL (-0.61; 0.5)

Veg (0.29; 0.2) + CC (0.003; 0.4) + GC (-0.09; 0.7) + LL (-1.13; 0.6)

Veg (-0.01; 0.3) + CC (-0.10; 0.4) + Veg*CC (0.39; 0.7)

0

0.63

0.74

0.94

1.44

1.81

2.36

2.54

2.94

3.79

3.92

4.17

4.93

1

0.73

0.69

0.62

0.49

0.40

0.31

0.28

0.23

0.15

0.14

0.12

0.09

0.19

0.14

0.13

0.12

0.09

0.08

0.06

0.05

0.04

0.03

0.03

0.02

0.02

220

(R) July GC (2.36; 1.0) + LL (1.93; 0.7)

Veg (0.54; 0.2)

LL (2.26; 0.7)

Veg (0.52; 0.3) + LL (3.45; 2.1) + Veg*LL (-2.72; 2.3)

Veg (0.51; 0.2) + GC (4.56; 3.0) + Veg*GC (-3.76; 3.3)

Veg (0.36; 0.4) + GC (5.08; 2.9) + LL (3.78; 2.0) + Veg*GC (-3.67; 3.3) + Veg*LL (-2.61; 2.3)

GC (3.03; 1.1)

0

2.61

2.69

3.99

4.64

5.08

5.83

1

0.27

0.26

0.14

0.10

0.08

0.05

0.53

0.14

0.14

0.07

0.05

0.04

0.03

221

August Constant only (0.64; 0.1)

GC (0.79; 1.1)

Veg (0.10; 0.2)

LL (-0.07; 0.7)

GC (0.85; 1.2) + LL (-0.19; 0.7)

Veg (0.36; 0.3) + LL (0.47; 2.3) + Veg*LL (-1.62; 2.5)

Veg (0.12; 0.2) + GC (2.68; 3.3) + Veg*GC (-2.54; 3.6)

Veg (0.49; 0.4) + GC (2.77; 3.3) + LL (0.65; 2.3) + Veg*GC (-3.31; 3.7) + Veg*LL (-1.96; 2.6)

0

1.71

1.82

2.18

3.90

5.40

5.75

9.49

1

0.43

0.40

0.34

0.14

0.07

0.06

0.01

0.41

0.18

0.17

0.14

0.06

0.03

0.02

0.004

222

September Constant only (1.20; 0.1)

GC (-1.19; 1.2)

Veg (0.09; 0.2)

LL (-0.27; 0.8)

CC (-0.03; 0.4)

Veg (0.14; 0.2) + GC (-6.74; 3.5) + Veg*GC (5.44; 3.9)

GC (-1.44; 1.3) + CC (-0.21; 0.5)

GC (-1.15; 1.3) + LL (-0.11; 0.8)

Veg (0.76; 0.4) + GC (-3.14; 1.6) + CC (0.24; 0.6) + LL (-1.90; 1.2)

LL (-0.42; 0.9) + CC (-0.15; 0.5)

CC (-0.33; 0.5) + GC (-1.43; 1.3) + LL (-0.40; 0.9)

Veg (0.28; 0.4) + LL (-1.10; 2.5) + Veg*LL (-0.02; 2.8)

Veg (0.25; 0.6) + CC (0.28; 0.7) + Veg*CC (-0.28; 1.5)

0

1.27

1.98

2.07

2.19

2.36

3.30

3.51

3.92

4.24

5.44

5.62

6.39

1

0.53

0.37

0.36

0.34

0.31

0.19

0.17

0.14

0.12

0.07

0.06

0.04

0.27

0.14

0.10

0.10

0.10

0.08

0.05

0.05

0.04

0.03

0.02

0.02

0.01

223

Bee (A) September CC (-4.83; 0.9) + LL (-4.05; 1.8) + GC (-5.45; 2.7)

Veg (0.78; 0.6) + CC (-3.89; 1.1) + GC (-6.61; 2.9) + LL (-5.58; 2.1)

LL (-4.72; 1.8) + CC (-4.74; 0.9)

GC (-6.43; 2.8) + CC (-4.11; 0.8)

CC (-3.94; 0.9)

Veg (0.03; 1.1) + CC (-4.44; 1.4) + Veg*CC (-1.82; 3.1)

0

0.94

0.97

2.16

4.27

7.60

1

0.63

0.62

0.34

0.12

0.02

0.37

0.23

0.23

0.13

0.04

0.008

224

(R) September Constant only (0.50; 0.1)

GC (1.02; 0.8)

CC (-0.32; 0.3)

Veg (0.11; 0.1)

LL (0.45; 0.5)

GC (0.77; 0.9) + CC (-0.22; 0.3)

GC (0.91; 0.8) + LL (0.32; 0.5)

Veg (0.13; 0.2) + GC (4.14; 2.3) + Veg*GC (-3.93; 2.6)

LL (0.18; 0.6) + CC (-0.27; 0.3)

Veg (-0.40; 0.4) + CC (-0.55; 0.4) + Veg*CC (1.31; 1.0)

Veg (-0.12; 0.2) + LL (-1.74; 1.6) + Veg*LL (2.49; 1.8)

CC (-0.17; 0.3) + GC (0.77; 0.9) + LL (0.17; 0.6)

Veg (-0.18; 0.3) + GC (1.16; 1.1) + CC (-0.30; 0.4) + LL (0.51; 0.8)

0

0.62

0.84

1.39

1.45

2.34

2.51

2.86

3.02

3.61

4.03

4.59

6.53

1

0.73

0.66

0.50

0.48

0.31

0.29

0.24

0.22

0.16

0.13

0.10

0.04

0.21

0.15

0.14

0.10

0.10

0.06

0.06

0.05

0.05

0.03

0.03

0.02

0.007

Veg = vegetation type; CC = canopy cover; LL = leaf litter; GC = ground cover. Model: parameter estimates for each predictor with SE in parentheses; Δi AICc = difference in AICc values, relative to the highest-ranked model; L(gi) = relative likelihood of the model; wi = Akaike weight for the model