this thesis is presented for the degree of doctor of

Effects of urbanization on remnant woodlands

Cristina E. Ramalho

MSc, PhD candidate

This thesis is presented for the degree of Doctor of Philosophy at

The University of Western Australia

School of Plant Biology

2012

i

Summary

This study aims to understand the effects of contemporary urbanization on remnant

plant communities. In growing cities worldwide, urbanization is leading to the rapid loss

of natural vegetation and its fragmentation into small and isolated urban remnants. Not

only is this phenomenon fairly recent, embodying the major urban transition the world

is currently experiencing and that gained momentum in the last 10-20 years, but also its

spatial extent is unprecedented. Although there is growing awareness of the impacts of

urbanization on biodiversity, how plant communities in these newly formed urban

remnants change over time following landscape fragmentation is not well-understood.

Yet, as cities sprawl, the importance of these remnant ecosystems for conservation of

biodiversity and provision of ecosystem services cannot be overlooked.

I studied 30 remnant Banksia woodlands in the rapidly urbanizing city of Perth, located

in the south-western Australian global biodiversity hotspot. First, I developed a

conceptual framework supporting the study of the effects of contemporary urbanization

on remnant ecosystems (Chapter 2). This framework highlights the importance of

considering the temporal dynamics of landscape change, land-use history, and the main

environmental drivers. Second, I attempted to identify the main driving factors of plant

community change in the remnant Banksia woodlands of the Perth Metropolitan Area. I

considered a comprehensive set of factors characterizing the landscape fragmentation

dynamics (current and past remnant area and landscape connectivity, time since

isolation and since urbanization, trajectories of landscape change), remnant local

environmental conditions (soil nutrient status and litter depth), and disturbance regimes

(fire, grazing, and human activities). Using an array of analytical methods, including

structural equation modelling, generalized mixed effect modelling, and fourth-corner

ii

analysis, I analysed three different aspects of the plant community. Initially, I examined

species richness and species abundance (Chapter 3). Then, I looked at community

functional composition, with a focus on growth form, seed dispersal, pollination

syndrome, nutrient acquisition, and regeneration strategy (Chapter 4). Finally, I

examined the canopy composition, density, and regeneration (Chapter 5). I then

summarized the results obtained in the thesis and put them in a broader context (Chapter

6).

The main results of the study are that: (1) richness and abundance of woody species

were higher in historically large remnants but lower in those located in the rural city

fringes; (2) richness of native herbaceous species declined with time since isolation,

mainly in the smaller remnants (1-5 ha), and with increasing soil organic C; (3)

increased abundance of non-native herbaceous species and increased litter depth, mostly

in the smaller remnants, were the main factors associated with the decline in the

abundance of native herbaceous species; (4) abundance of non-native herbaceous

species was lower in remnants strongly grazed by native herbivores, but higher in

remnants grazed by non-native herbivores. From a functional perspective, the study

showed that: (5) in comparison with similarly sized remnants that were recently

fragmented, the smaller (1-5 ha) and older remnants had: (6) lower relative abundance

of shrubs and higher relative abundance of trees and sedges; (7) lower relative

abundance of understorey plant species pollinated by insects and dispersed after animal

ingestion, and higher relative abundance of understory plant species dispersed and

pollinated by wind; (8) a different overstorey composition, with a higher relative

abundance of ectomycorrhizal tree species (Eucalyptus spp.) in detriment of non-

mycorrhizal root-clustered tree species (Banksia spp.). Finally, a study of the canopy

indicated that: (9) high fire frequency and perhaps other non-measured factors, such as

iii

declining ground water levels, may contribute to the lower density of live trees and

recruits of the co-dominant Banksia attenuata in the older urban remnants.

Overall, this study indicates that in a city with a recent development history, where

urbanization is an important driver of habitat fragmentation, plant communities are in a

slow trajectory of change. A “time lag effect”, mostly in the smaller remnants, is likely

to encapsulate the effects of fragmentation on the colonization-extinction dynamics and

on the disturbance regimes, and indicate the trajectory in which the plant communities

will continue to change over time.

v

Table of contents

SUMMARY .......................................................................................................................................... I

TABLE OF CONTENTS .......................................................................................................................... V

ACKNOWLEDGEMENTS ...................................................................................................................... IX

STATEMENT OF CANDIDATE CONTRIBUTION ........................................................................................ X

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

1.1 BACKGROUND .............................................................................................................................. 1

1.2 BANKSIA WOODLANDS IN THE PERTH METROPOLITAN AREA ........................................................... 3

1.3 KEY RESEARCH QUESTIONS ........................................................................................................... 5

1.4 THESIS OUTLINE ........................................................................................................................... 5

2. A DYNAMIC APPROACH IN THE STUDY OF THE EFFECTS OF URBANIZATION ON

REMNANT PLANT COMMUNITIES ............................................................................................... 7

2.1 CHALLENGES FOR URBAN ECOLOGY IN A RAPIDLY URBANIZING WORLD .......................................... 7

2.2 CONCEPTUAL FRAMEWORKS IN URBAN ECOLOGY .........................................................................11

2.2.1 The Urban-to-Rural Gradient Framework ...........................................................................11

2.2.2 Other frameworks ...............................................................................................................13

2.3 LIMITS TO CURRENT APPROACHES ................................................................................................14

2.3.1 Linear gradients do not fit with the characteristics of contemporary cities ...........................14

2.3.2 Simplification of the set of intervening drivers .....................................................................16

2.3.3 Lack of a temporal perspective ............................................................................................17

2.4 THE IMPORTANCE OF A TEMPORAL PERSPECTIVE ...........................................................................18

2.4.1 Land-use legacies................................................................................................................19

2.4.2 Time-lagged responses to fragmentation ..............................................................................19

2.5 TOWARDS AN EMERGING FRAMEWORK IN URBAN ECOLOGY ..........................................................21

2.6 APPLICATION TO PLANNING, MANAGEMENT, AND RESTORATION IN CONTEMPORARY CITIES............25

2.7 CONCLUDING REMARKS ...............................................................................................................26

vi

3. DELAYED EFFECTS OF FRAGMENTATION ON PLANT SPECIES RICHNESS AND

ABUNDANCE IN REMNANT WOODLANDS OF A RAPIDLY URBANIZING BIODIVERSITY

HOTSPOT .......................................................................................................................................... 29

3.1 INTRODUCTION ........................................................................................................................... 30

3.2 METHODS ................................................................................................................................... 34

3.2.1 Study area .......................................................................................................................... 34

3.2.2 Remnant selection and sampling design .............................................................................. 36

3.2.3 Vegetation survey ............................................................................................................... 37

3.2.4 Landscape and local factors ............................................................................................... 38

3.2.5 Data analysis...................................................................................................................... 41

3.3 RESULTS .................................................................................................................................... 42

3.3.1 Interactive effect of time since isolation and remnant area on species richness .................... 43

3.3.2 Effects of landscape and local factors on species richness ................................................... 43

3.3.3 Disentangling the effect of landscape and local factors on species richness and abundance . 46

3.4 DISCUSSION ................................................................................................................................ 48

3.4.1 Delayed effects of fragmentation on urban remnant vegetation ............................................ 48

3.4.2 Effects of landscape connectivity and agrarian land uses..................................................... 49

3.4.3 Natural and anthropogenic disturbances ............................................................................. 50

3.4.4 Final remarks ..................................................................................................................... 53

4. PLANT FUNCTIONAL COMPOSITION OF REMNANT WOODLANDS IN A RAPIDLY

URBANIZING BIODIVERSITY HOTSPOT ................................................................................... 55

4.1 INTRODUCTION ........................................................................................................................... 56

4.2 METHODOLOGY .......................................................................................................................... 60

4.2.1 Study area and study remnants............................................................................................ 60

4.2.2 Landscape and local factors ............................................................................................... 60

4.2.3 Vegetation survey ............................................................................................................... 60

4.2.4 Plant functional traits ......................................................................................................... 60

4.2.5 Data analysis...................................................................................................................... 62

4.3 RESULTS .................................................................................................................................... 63

4.3.1 Functional composition of the overstorey plant community .................................................. 64

4.3.2 Functional composition of the understorey plant community ................................................ 66

vii

4.4 DISCUSSION ................................................................................................................................69

5. DRIVING FACTORS OF TREE MORTALITY AND REGENERATION IN

MEDITERRANEAN URBAN WOODLANDS ..................................................................................77

5.1 INTRODUCTION ...........................................................................................................................78

5.2 METHODOLOGY ..........................................................................................................................81

5.2.1 Study area ...........................................................................................................................81

5.2.2 Remnant selection and sampling design ...............................................................................81

5.2.3 Canopy survey ....................................................................................................................81

5.2.4 Environmental factors .........................................................................................................81

5.2.5 Data analysis ......................................................................................................................82

5.3 RESULTS .....................................................................................................................................83

5.4 DISCUSSION ................................................................................................................................86

6. CONCLUSIONS .............................................................................................................................93

6.1 KEY RESULTS ..............................................................................................................................93

6.2 THE EFFECTS OF URBANIZATION ON REMNANT PLANT COMMUNITIES – REFLECTIONS IN A BROADER

CONTEXT ..........................................................................................................................................97

6.3 LIMITATIONS OF THE STUDY ...................................................................................................... 100

6.4 FUTURE DIRECTIONS .................................................................................................................. 101

6.4.1 Urbanization as a set of filters ........................................................................................... 101

6.4.2 A temporal perspective in the study of rapidly urbanizing landscapes ................................. 103

6.4.3 Supporting urban planning, management, and restoration of urban remnants in a rapidly

urbanizing world ....................................................................................................................... 104

LITERATURE CITED ..................................................................................................................... 105

APPENDICES .................................................................................................................................. 135

ix

Acknowledgements

I am grateful to my main supervisor Richard Hobbs for his continuous support from the

beginning to the end of this journey. I am also thankful to my co-supervisors Etienne Laliberté

and Pieter Poot for their help and insightful input on the data chapters.

I am grateful to Mark Brundrett, Kingsley Dixon, Neil Gibson, Ray Froend, Michael Hammond,

Hans Lambers, Alex Chapman, Erik Veneklaas, Neil Enright and Deanna Rokich, for their help

on the plant trait data collection and discussions about the local flora. I am also thankful to

Lauren Hallett, Michael Perring, Kris Hulvey, and two anonymous reviewers for insightful

comments on Chapter 2.

I am grateful to Paul Liebich, Juan Garibello, Graham Zemunik, and Sutomo for their help on

fieldwork and to Phil Ladd and Bronwen Keighery for their help with plant identification.

I express my gratitude to the ERIE group for providing me with such an international, friendly,

and stimulating working environment: Christine Allen, Heather Gordon, Hilary Harrop-

Archibald, Kris Hulvey, Joanna Burgar, John Dwyer, Juan Garibello, Keren Raiter, Lauren

Hallett, Letícia Garcia, Lisa Denmead, Lori Lach, Maggie Triska, Mandy Trueman, Martha

Orozco, Michael Craig, Michael Perring, Nancy Shackelford, Pat Kennedy, Peter Grose,

Rachel Standish, Rebecca Parsons, Sue Yates, Sutomo, Tim Morald.

I express my gratitude to my dearest friends who have made my life much more special during

the PhD: Amélia Martins-Loução, Ana Francisco, Ann Hamblin, Annie Hurwitz-Forshaw,

Christine Soulier, Claudia Fulgencio, Christian Hermann, Danny Thies, David Hurwitz-

Forshaw, Emmaclair Bussel, Gillian Henderson, Helen Heppingstone, Jean-Paul Orsini, Joana

Matos, Julio Martín, Lesley Shaw, María Calviño Cancela, Maria Grade, Matthias Boer,

Natasha Stone, Nicole Pinnel, Patrick Lienin, Paul Audin, Pedro Bettencourt, Telma Esteves.

I am grateful to my dearest family for their unconditional love and support.

I want to acknowledge the generous financial support of the Portuguese National Science

Foundation (Fundação para a Ciência e a Tecnologia), who provided my doctoral scholarship;

and the School of Plant Biology, University of Western Australia, who has funded the study.

Finally, I want to thank the Banksia woodlands. They have opened my eyes to the amazing

natural history of south-western Australia.

x

Statement of candidate contribution

This thesis includes material from one paper that has already been published and which

is jointly authored:

- Ramalho, C.E., Hobbs, R.J. (2012). Time for a change: dynamic urban ecology.

Trends in Ecology and Evolution, 27 (3): 179-188 (Chapter 2)

CER (75%) developed the concepts and ideas in the paper and wrote the initial draft.

RJH (25%) contributed to discussions on concepts, some text sections, and final text

editing.

The thesis also contains material from three other papers prepared for publication

(Chapters 3-5), all of which have been co-authored. CER (75%) is the primary author on

all the papers and conducted the fieldwork, data analysis, and wrote the initial drafts.

Co-authors EL, PP, and RJH (25%) have provided valuable advice on data analysis,

contributed with ideas, and final text editing. Because the thesis is constructed as a

series of papers to be submitted for publication, there is some degree of repetition in the

introductory sections of some of the chapters.

Cristina E. Ramalho

Prof Richard J. Hobbs Dr Pieter Poot Dr Etienne Laliberté

1

1. INTRODUCTION

1.1 Background

We live in a rapidly urbanizing world. As recently as 1950, only 30% of the world

population was urbanized. Today, more than 50% of the world population lives in urban

areas and this number is expected to reach 70% by 2050. Similarly, by 1950 the number

of cities with > 1 million inhabitants was below 100, whereas that number has now

reached 500. These figures, among others recently published by the United Nations

(UNFPA, 2007; UN, 2010) reflect an unprecedented urban transition, with

characteristics that are different from any other moment in history. An important feature

of contemporary urbanization is that cities are no longer compact but increasingly

dispersed and expansive (UNFPA, 2007; Seto et al., 2010). Indeed, they sprawl in

fractal or spider-like configurations (Makse et al., 1995; Batty, 2008), interspersing with

rural and protected areas across large regions (Schneider and Woodcock, 2008;

Wittemyer et al., 2008; Radeloff et al., 2010). Consequently, contemporary urbanization

is a major driving force of habitat fragmentation and species endangerment (Czech et

al., 2000; Robinson et al., 2005; Burgman et al., 2007). Given the unprecedented pace

of current urban expansion, there may be a mismatch in public and scientific perception

of the extent of this phenomenon (Borgström et al., 2006). However, as cities sprawl

and leave behind a “patchwork” of scattered and isolated remnant vegetation fragments,

the importance of those remnants for conservation of biodiversity and provision of

ecosystem services cannot be overlooked.

A range of studies has described patterns of plant species distribution and, less

frequently, plant functional traits, along urban-to-rural gradients (e.g., Guntenspergen

2

and Levenson, 1997; Bastin and Thomas, 1999; Godefroid and Koedam, 2003;

Williams et al., 2005; Celesti-Grapow et al., 2006; Guirado et al., 2006; Duguay et al.,

2007; Hahs and McDonnell, 2007; Gavier-Pizarro et al., 2010; Vallet et al., 2010).

Other studies have also described patterns of variation in ecosystem processes,

including leaf litter decomposition, nitrogen, and carbon cycling, microbial biomass,

and seedling establishment (Pouyat et al., 1997; Carreiro et al., 1999; Pouyat and

Turechek, 2001; Zhu and Carreiro, 2004; Kaye et al., 2005). These studies among

others have laid the foundation of what is known about the effects of urbanization on

remnant ecosystems (McKinney, 2008).

Certain features of rapidly urbanizing cities have major ecological implications that

have not been properly acknowledged in previous studies of urban remnants. First,

urbanizing landscapes have a changing nature. In other words, their landscape

configurations have been changing significantly in the last few decades, most

predominantly in the last 10-20 years, and will likely continue to change over time, as

urban growth occurs through infilling. Second, arising from this phenomenon is the fact

that patches of remnant vegetation vary in their degree of fragmentation and

urbanization age. These measures might be highly correlated in cities expanding largely

into native vegetation, but uncorrelated if cities are expanding over previously

fragmented landscapes (e.g. agricultural areas). Importantly, given that current urban

expansion is increasingly non-linear and dispersive, those measures do not necessarily

follow a predictive urban-to-rural gradient as measured by the distance of a fragment to

the city centre. Third, urban areas are more interspersed with rural and protected areas,

and the boundaries between “urban” and “rural” are less sharp than before (Seto et al.,

2010). In other words, rapidly urbanizing landscapes are not only “urban” but they

embed fragments of other land uses (e.g., industry, agriculture, forestry).

3

Most studies describing plant species distribution along urban-to-rural gradients use a

static approach that does not consider the temporal dynamics of landscape change and

the effects of past land-use legacies (exceptions include McKinney, 2002; Hahs et al.,

2009). Furthermore, they have also relied on the use of urban-to-rural geographic

transects either alone (McDonnell and Pickett, 1990) or coupled with aggregated

urbanization variables to represent the urban gradient as explanatory variables of

ecological variation (McDonnell and Pickett, 1990; Luck and Wu, 2002; McDonnell

and Hahs, 2008). Given the complex combination of factors in urban areas (Liu et al.,

2007), a more comprehensive approach considering the main environmental filters, the

temporal dynamics of landscape change, and land-use history, is likely to further

enhance current understanding of the effects of urbanization on remnant plant

communities.

1.2 Banksia woodlands in the Perth Metropolitan Area

This study aimed to contribute to the understanding of how urbanization affects remnant

plant communities in rapidly urbanizing landscapes, with an emphasis on tackling some

of the limitations of previous studies. To do so, the remnant Banksia woodlands of the

Perth Metropolitan Area (Western Australia) were adopted as a case study. Perth is a

relatively young and rapidly growing city that is expanding largely into intact native

vegetation. Indeed, the city was founded by Europeans in 1829 (Kennewell and Shaw,

2008). Prior to that, indigenous practices of burning, hunting, and gathering shaped the

region for a long period (20,000-40,000 years) (Seddon, 1972). From the 1960’s

onwards, the city has witnessed a steep urban growth fuelled by a mining boom that

now drives the state’s economy (Kennewell and Shaw, 2008; Weller, 2009). Perth’s

current population of 1,659,000 is estimated to reach 3 million by 2050 (Australian

4

Bureau of Statistics, 2011). The city presently extends over 120 km along the coast and

covers 100,000 ha (Weller, 2009). Furthermore, it is dominated by 1–2 storey houses at

an average density of six houses per hectare (Weller, 2009), and with only a few,

localized industrial and high-density building areas. These patterns of urban sprawl raise

several environmental concerns associated with the high demand for land and

dependence on private cars (Kennewell and Shaw, 2008). Importantly, urbanization has

been the main driver of habitat loss and fragmentation in the region (Burgman et al.,

2007). This is of particular concern given the location of Perth in the south-western

Australian Floristic Region, a global biodiversity hotspot with exceptionally high levels

of floristic diversity, endemism, and species geographic turnover (Myers et al., 2000;

Hopper and Gioia, 2004).

Thirty remnant Banksia woodlands located in a radius of 30 km from the Perth city

centre were selected for the study. Banksia woodlands are Mediterranean fire-prone

plant communities with an open canopy dominated by Banksia attenuata and B.

menziesii (Proteaceae), and a species-rich understorey dominated by sclerophyllous

shrubs. Between 2008 and 2009, the 30 remnants were surveyed for their plant

community composition (species abundance), canopy (tree density, mortality, and

regeneration), and main perceived vectors of disturbance. The latter included human

activities (e.g., trampling, waste disposal, and soil physical disturbance), grazing, and

elevated soil nutrient levels. Further information about the remnants, including fire

history, current and past spatial configurations and land uses, time since isolation and

time since urbanization, was collected using a Geographic Information System that was

assembled for the study. The data collected were then analysed in a number of ways to

address the main research questions of the study.

5

1.3 Key research questions

The key research questions of the study include:

What are the main landscape and local factors driving changes in the species

richness and abundance of remnant plant communities in rapidly expanding

cities?

What are the main landscape and local factors driving changes in the plant

functional composition of remnants plant communities in rapidly expanding

cities? Are there plant traits being consistently favoured or disfavoured as

remnants age in the urban matrix?

What are the main landscape and local factors influencing tree mortality and

regeneration in remnant woodlands of a rapidly urbanizing city?

To what extent are the effects of contemporary urbanization currently observable

in urban remnant woodlands? Is there an interactive effect between remnant area

and time since isolation determining stronger and quicker changes in the plant

community of the smaller and older urban remnants?

1.4 Thesis outline

Including the current introductory chapter, this thesis is structured into six main

chapters:

Chapter 1 presents the thesis background, the case study analysed, key

research questions, and thesis outline.

Chapter 2 reviews how urbanization has been assessed in urban ecological

studies, and proposes a framework for the study of the effects of contemporary

urbanization on remnant ecosystems.

6

Chapter 3 investigates the effects of a comprehensive set of landscape and

local urbanization-related factors on species richness and abundance.

Chapter 4 investigates the effects of the same factors on the plant community

functional composition, with a focus on growth form, seed dispersal mode,

pollination syndrome, nutrient acquisition, and regeneration strategy.

Chapter 5 investigates what factors influence mortality, density, and

regeneration of the four most abundant tree species.

Chapter 6 summarizes the results obtained in the thesis, discusses their

broader relevance, identifies limitations of the study, and highlights future

directions of research.

7

2. A DYNAMIC APPROACH IN THE STUDY OF THE EFFECTS OF

URBANIZATION ON REMNANT PLANT COMMUNITIES1

Contemporary cities are expanding rapidly in a spatially complex, non-linear manner.

However, this form of expansion is rarely taken into account in the way that

urbanization is classically assessed in ecological studies. An explicit consideration of

the temporal dynamics, although frequently missing, is crucial in order to understand

the effects of urbanization on biodiversity and ecosystem functioning in rapidly

urbanizing landscapes. In particular, a temporal perspective highlights the importance

of land-use legacies and transient dynamics in the response of biodiversity to

environmental change. Here, we outline the essential elements of an emerging

framework for urban ecology that incorporates the characteristics of contemporary

urbanization and thus empowers ecologists to understand and intervene in the planning

and management of cities.

2.1 Challenges for urban ecology in a rapidly urbanizing world

Not only is the world experiencing an unprecedented urban transition (UNFPA, 2007;

UN, 2010), but contemporary urbanization also differs markedly from historical patterns

of urban growth (Seto et al., 2010) (Box 2.1), thus imprinting a unique signature on

contemporary cities (Figure 2.1). Indeed, such cities are largely young urban landscapes

that have expanded rapidly over the course of the major urban transition that started in

1950 and that has accelerated steeply over the past 10-20 years (UNFPA, 2007; Seto et

1 This chapter has been published as: Ramalho, C. E and Hobbs R. J. (2012). Time for a change: dynamic urban ecology. Trends in Ecology and Evolution, 27 (3): 179-188

8

al., 2010; UN, 2010). Importantly, contemporary cities are increasingly expansive and

dispersed landscapes (Seto et al., 2010; Seto et al., 2011), which grow and age in a

spatially complex, non-linear manner (Batty, 2008). Consequently, they display

multifaceted patterns of variable density across space and time, in which high density

built-up areas can be finely interspersed with lower density, rural, and natural areas

(Alberti, 2008; Seto et al., 2010). By contrast, historically developed cities are

contained areas that grew slowly, over several centuries or decades, through concentric

and compact rings of development (Seto et al., 2010). Contemporary urbanization has

major implications for the ecology of cities, requiring ecologists to acknowledge the

phenomenon actively in terms of the ways they intervene in, and study, cities.

Figure 2.1 Contemporary cities. Aerial perspectives of Chicago (a, b) and Houston, USA (c, d) illustrating

how contemporary urbanization imprints a unique signature on cities. Contemporary cities are largely

young and rapidly evolving urban landscapes that have expanded dramatically over the past few decades,

during the second major world urban transition. These cities no longer have a compact development, but

instead are highly expansive and dispersed, sprawling in fractal or spider-like configurations (Batty,

2008), and embedding functioning or decaying fragments of other land uses (e.g., agriculture, forestry or

remnant vegetation) in the rapidly changing matrix. The complex spatial patterns of urban growth reflect

not only past landscape configurations, but also current socioeconomic and political processes, such as

planning, transportation costs, agglomeration economies, and market prices (Seto et al., 2010).

Reproduced with permission from R. Hobbs (a, b) and C. E. Ramalho (c, d).

9

Box 2.1 A rapidly urbanizing world

Since 2008, for the first time, more than half of the human population of the world (3.3 billion),

lives in urban areas and this number is expected to reach 5 billion by 2030 (UN, 2010). This

figure reflects an unprecedented urban transition, with characteristics that are different from any

other moment in history. In a thorough recent review (Seto et al., 2010), it was shown that

contemporary urbanization markedly differs from historical patterns of urban growth in terms of

scale, rate, location, and form. First, the scale and rate of urban expansion, both in terms of

population growth and land-cover change, are extraordinary. For instance, between 2000 and

2030 middle-sized cities with populations of 500 000 to 1 million are expected to triple their

area (UNFPA, 2007). Second, the location of urbanization is shifting. Indeed, whereas the first

urban transition (1750-1950) took place in Europe and North America, increasing their urban

population from 15 million to 423 million, the second urban transition (1950-2030) is

happening largely in Africa and Asia and will increase their urban population from 309 million

to 3.9 billion in only 80 years (UNFPA, 2007; Montgomery, 2008). By 2030, these countries

will contain 80% of world urban population. Third, the shape of the cities has changed. Whereas

historical cities were contained and well defined areas that grew through concentric rings

surrounding a dense urban core, contemporary cities are no longer sharply defined and are

increasingly dispersed and expansive (Seto et al., 2010). Furthermore, the patterns of urban

sprawl differ between countries. Indeed, in places such as the USA and Australia,

suburbanization is predominant and consists of single-family residential development. In

developing countries and some European cities, sprawl occurs predominantly through peri-

urbanization, a more disordered development that expands along urban corridors spreading out

from metropolitan regions and incorporating small towns and rural areas (Seto et al., 2010).

As cities expand, protected areas that today are currently outside city boundaries will

soon become embedded in urban landscapes (McDonald et al., 2008; Wittemyer et al.,

2008; Radeloff et al., 2010; Seto et al., 2011). Furthermore, other natural areas and

previously managed land with conservation value (e.g., old fields) will be largely

reduced to small and scattered urban remnants. Whereas cities were previously

relatively confined spaces and therefore conservation of remnant ecosystems within

their boundaries was not a priority, this is no longer the reality. In fact, the conservation

of urban remnant ecosystems will become increasingly important for several reasons.

First, especially in areas with high beta-diversity, remnants provide the only remaining

10

habitat for many species (Miller and Hobbs, 2002). Second, they provide ecosystem

services (e.g., water infiltration, micro-climatic amelioration, sequestration of air

pollutants, recreation, and aesthetics) that improve the urban environment and enhance

the wellbeing and quality of life of urban dwellers (Bolund and Hunhammar, 1999;

Fuller et al., 2007; Jim and Chen, 2009). Third, urban remnants are the primary

connection that many humans have to the natural world (Sanderson and Huron, 2011).

Preventing the extinction of this experience (Miller, 2005) is important for conservation

far beyond city boundaries (Dunn et al., 2006).

Urban ecological research is largely framed by a conceptual approach that assumes that

urbanization and its induced environmental changes decrease in a linear gradient from

the core to the city fringes (McDonnell and Pickett, 1990). This assumption, as well as

oversimplifying urban environments (Theobald, 2004; Alberti, 2008), does not fit with

the non-linear and complex growth of contemporary cities. Equally important, a static

approach neglecting the young and rapidly evolving nature of those landscapes (and

consequent ecological implications) is predominant across current urban ecology

frameworks. This might result from a slow recognition of the unprecedented spatial and

temporal scale of contemporary urbanization (Borgström et al., 2006). Regardless of its

cause, this mismatch has major consequences for the scope of urban ecological research

and calls for an urgent revision in which urbanization is assessed in ecological studies.

Here, we review how urbanization is evaluated in ecological studies. We identify key

drawbacks of current conceptual frameworks are identified, emphasizing the misleading

assumptions of linear variation in urbanization intensity and age, the simplification of

the set of intervening drivers, and the lack of a temporal approach. We then propose an

emergent framework for urban ecology: the Dynamic Urban Framework. This

11

incorporates an explicit temporal perspective that considers land-use legacies and time-

lagged ecological responses to ongoing environmental change. Furthermore, it includes

a conceptual and analytical structure where relationships between intervening drivers

can be analysed in a mechanistic manner. Here, it focuses on remnant ecosystems, but is

extendable to other components of the urban environment. Finally, the framework can

be incorporated or used in conjunction with other conceptual frameworks with a

stronger multidisciplinary focus (Grimm et al., 2000; Alberti et al., 2003). It is time for

a change in the way in which ecologists view, study, and intervene in cities, so that they

can have a more active and positive role in the planning and management of the places

in which most humans now live.

2.2 Conceptual frameworks in urban ecology

2.2.1 The Urban-to-Rural Gradient Framework

The urban-to-rural gradient approach (McDonnell and Pickett, 1990) has framed most

ecological studies analyzing the effects of urbanization on biodiversity and ecosystem

functioning (Theobald, 2004; McKinney, 2008; Pouyat et al., 2008). For example, it

aided in the understanding of how species richness varies across urban-to-rural

gradients (Chace and Walsh, 2006; McKinney, 2008) and in response to important

urbanization drivers, such as population density (Luck, 2007). This framework views

cities predominantly as monocentric or sometimes polycentric agglomerations that grow

through concentric rings surrounding a dense urban core (McDonnell and Hahs, 2008).

Most importantly, the framework assumes that urbanization and its induced

environmental changes vary along linear gradients between the urban core and the

peripheral rural matrix (Alberti, 2008). These include changes in land cover, species

12

assemblages, the chemical and physical environment, and disturbance regimes

(McDonnell and Hahs, 2008).

Framed by the urban-to-rural gradient, urbanization is depicted and assessed in

ecological studies in two main ways (Theobald, 2004; McDonnell and Hahs, 2008). A

first group of studies simply uses broad zoning categories that are defined subjectively

based on the general landscape context (Garden et al., 2010) or along a geographical

transect (Magura et al., 2010). Such studies compare responses between sites located in,

for instance, urban, suburban, and rural areas (Koerner and Klopatek, 2010); urban,

rural and natural areas (Garden et al., 2010), or city centre, city edge and peri-urban

areas (Hedblom and Söderström, 2010) (Box 2.2). Alternatively, linear distance to the

city centre has been used as a precise measure of the gradient (Williams et al., 2005). A

second group of studies combines gradient analysis with landscape metrics (Luck and

Wu, 2002; Hahs and McDonnell, 2006) and/or land-use types (Blair, 1996). In the first

case, census, cartography, and remote-sensing data are used to quantify socio-economic,

land cover, land use, and built infrastructure variables in or around the study sites.

These variables are used individually or aggregated as proxies to characterize the degree

of urbanization. Commonly used metrics measure population density (Hahs and

McDonnell, 2006), income (Loss et al., 2009; Lerman and Warren, 2011), percentage of

impervious surface (Lu and Weng, 2006), housing (Germaine and Wakeling, 2001) and

road density (Williams et al., 2005). This approach has recently featured in studies

aiming to define standardized measures of urbanization to be used in comparative

studies (McDonnell and Hahs, 2008; du Toit and Cilliers, 2011).

13

2.2.2 Other frameworks

Other conceptual frameworks have been proposed in urban ecology, whose use has been

restricted to a few specific case studies (Pickett et al., 1997; Grimm et al., 2000; Alberti

et al., 2003). These frameworks are strongly based on the integration of social and

environmental sciences. Importantly, they have an ecosystem focus, exploring the links

between human and biophysical drivers, patterns, and processes, to understand the

relationships between urbanization and ecosystem functioning. The Human Ecosystem

Model (Machlis et al., 1997; Pickett et al., 1997) is strongly rooted in social sciences

and, together with the hierarchical patch dynamics framework (Wu and Loucks, 1995)

and watershed models (Band and Moore, 1995), provides the conceptual and analytical

core for the urban Long-Term Ecosystem Research (LTER) projects in Baltimore and

Central Arizona-Phoenix (Pickett et al., 2001; Pickett and Cadenasso, 2006). Other

frameworks include the Human Modification Framework (Theobald, 2004), the

Multidimensional Biocomplexity Framework (Pickett et al., 2005), and the emerging

framework based on the LTER Baltimore Ecosystem Study (Pickett et al., 2008).

Box 2.2 Lost in translation

Several terms and urbanization measures originating from social sciences, geography, and urban

planning are widespread in the ecological literature and adopted in core research aspects, such

as study design. However, these are often used without a critical assessment of their ecological

meaning (McIntyre et al., 2000) or of whether they reflect the actual range of disturbance to

which ecosystems are exposed in the study context. Urban ecological research is hence wedded

to broad, vague terms and measures of urbanization, which impedes ecologists from gaining a

more mechanistic understanding of the ecology of cities. This problem might result from two

reasons. First, urban ecology has had its major methodological and conceptual development in

social sciences, geography, and urban planning, and has only recently emerged in mainstream

ecological research (Pickett and Cadenasso, 2006; Young, 2009). Consequently, relevant bodies

of knowledge from those fields might have been poorly transferred into ecological language and

focus. This has the pernicious consequence that recent advances in the other fields are little

14

recognized in urban ecological research. For instance, whereas the spatial and temporal

complexity of cities is comprehensively acknowledged in urban planning (e.g., Iacono et al.,

2008), ecological studies still approach those characteristics in a very rudimentary way. There is

therefore a gap between fields, and important considerations get lost in translation. Second, the

use of broad urbanization terms and measures has been promoted as a common platform for

data collection and integration across different fields and in comparative studies (Theobald,

2004; Tress et al., 2004). However, it is almost impossible to determine the definition of single

terms or a set of urbanization metrics that are universally applicable (McIntyre et al., 2000;

Andersson et al., 2009; Tavernia and Reed, 2009). The quest for integration in such a

multidisciplinary field is important, but must happen in parallel with the development of

specific and ecologically driven vocabulary, concepts, and theories.

2.3 Limits to current approaches

2.3.1 Linear gradients do not fit with the characteristics of contemporary

cities

Urban-to-rural gradient studies oversimplify cities (Theobald, 2004; Alberti, 2008).

Initially, this simplification was important in developing an understanding of these

highly complex human-modified ecosystems. However, the underlying assumption of a

linear gradient in the urbanization-induced environmental changes does not fit with the

spatial-temporal characteristics of contemporary urbanization. Indeed, the fact that

contemporary cities grow in a rapid, complex, non-linear, dispersed, and expansive

manner, means that urbanization intensifies and ages in patchy and complex spatial

patterns across the landscape, rather than in a linear gradient (Figure 2.1). Consequently,

the environmental or ecological conditions in one focal remnant patch depend not on its

position along the linear gradient, but on the characteristics of the neighbouring patches.

In a similar way, remnants closer to the city have not necessarily been isolated for

longer than remnants in rural areas, and remnants close to each other might have been

isolated for different lengths of time (Figure 2.1). This means that the use of categorical

15

or quantitative measures of geographical linear distance in urban ecological studies can

be highly ambiguous and misleading.

Remnant vegetationRemnant vegetation Urban matrixUrban matrix Sampling siteSampling siteAgricultural matrixAgricultural matrix

A

B

C

D

Current10 years ago20 years ago30 years ago

Figure 2.2 Urban growth is a dynamic process in space and time. Traditional approaches measuring the

degree of urbanization in study sites often neglect the temporal dynamics of landscape change, with only

the most recent spatial configuration and surrounding land uses taken into account. This approach provides only a snapshot in which all the dynamics that led to that captured spatial moment are neglected,

severely limiting understanding into the past and future. In this figure, the landscape context and spatial

configuration of four different remnants (a-d) are represented along a period of 30 years. A “snapshot”

approach taken at the current time would classify these remnants in the same category. However, their

trajectories of landscape change show that the intensity of exposure through edge effects to the

disturbance processes originating in the surrounding urbanized areas is different in the four cases. While

remnant (a) has been isolated for 30 years, with the same spatial configuration, remnant (b) has only

recently had its area reduced to the same size and, in the near past, was part of a much larger and

continuous remnant. This means that, in remnant (a), the sampling site has been highly exposed through

edge effects to the urban disturbance processes from the immediate vicinity, and probably is highly

degraded, unless it has been targeted by management and restoration efforts. In remnant (b), the exposure

to an edge is relatively recent, and communities in the sampling site might or might not already exhibit an altered composition and structure owing to the current spatial configuration. Whereas in (a) and (b) the

major driver of landscape fragmentation was urbanization, in remnants (c) and (d), the major driver was

agriculture. This means, first, that the isolation history of the remnant could be much older and, second,

that both (c) and (d) have been exposed for a long time to an agriculture matrix and its disturbance

processes, only recently being exposed to urbanization. Solely considering the current landscape context

means that land-use legacies from the surrounding agricultural matrix prior to the onset of urbanization

are missed.

16

2.3.2 Simplification of the set of intervening drivers

Urban-to-rural gradient studies using landscape metrics and/or urban land-use types can

partially capture some of the non-linear heterogeneity and complexity of cities.

Nevertheless, these studies still oversimplify urban environments, as they often “flatten”

several human and environmental drivers into a reduced number of aggregated variables

used in study design and data analysis (Alberti et al., 2003; Theobald, 2004; Alberti,

2008), although there are a few exceptions (Schlesinger et al., 2008; Loss et al., 2009;

Lerman and Warren, 2011). The aggregated representation of drivers does not fully

encapsulate the complex dynamics in urban ecosystems, because it neglects the role of a

broader set of drivers and their interactions affecting remnant biodiversity and

ecosystem functioning. These drivers include, for instance, landscape fragmentation,

disturbance regimes, local environmental conditions, and the features of the local

environment that are not affected by urbanization. Biotic responses to environmental

changes associated with urbanization might be masked if such factors are ignored

(Cuffney et al., 2010). In fact, the action and interaction of multiple drivers, including

those that are unique to cities, is responsible for different processes and dynamics of

disturbance (Borgström et al., 2006; Liu et al., 2007) that can even decouple

fundamental ecological mechanisms (Shochat et al., 2006). Predator–prey relationships

can break down because synanthropic predators become strongly subsidized by

anthropogenic resources (Rodewald et al., 2011). Urban species assemblages can also

be determined mainly by stochastic processes rather than by mechanisms such as

interspecific competition (Sattler et al., 2010; Shochat et al., 2010). For these reasons,

single or simple combinations of aggregated urbanization measures must be used with

caution, and an explicit quantification of the intervening drivers and their interactions is

required. The absence of such an approach limits the capacity to understand and forecast

the effects of urbanization on remnant biodiversity and ecosystem functioning, as well

17

as their combined effects with other global change drivers, such as climate change

(Grimm et al., 2008; Nelson et al., 2009). Moreover, it diminishes the ability to provide

ecologically derived guidelines for management and restoration (Miller and Hobbs,

2002; Pressey et al., 2007).

Other frameworks (Pickett et al., 1997; Grimm et al., 2000; Alberti et al., 2003) are

generally based on a comprehensive set of human and biophysical drivers. However, the

focus on social sciences, and ecosystem processes and functions rather than on

biodiversity dynamics has reduced the ecological utility of those frameworks and might

explain their lack of application in urban ecological studies. This simplification is not

problematic in the study of the human-created and highly managed habitats (e.g., parks,

lawns, or green roofs). However, it becomes relevant if applied to remnant ecosystems,

as aspects such as fragmentation and associated direct and indirect effects on

biodiversity are not properly accounted for.

2.3.3 Lack of a temporal perspective

Another major limitation of current conceptual frameworks is the absence of an explicit

temporal perspective. Although the importance of temporal dynamics is well recognized

in ecology (Strayer et al., 2006; Willis and Birks, 2006; Hastings, 2010; Jackson and

Sax, 2010), this has been incorporated less well into the study of cities. However,

urbanization age (Loss et al., 2009; Magle and Crooks, 2009; Park et al., 2010; Lerman

and Warren, 2011), time-lagged social factors (Pickett et al., 2008; Boone et al., 2009),

the development history of the cities (Hahs et al., 2009), and their agrarian legacies

(Lewis et al., 2006) have all been shown to be important factors in determining current

urban biodiversity and ecosystem patterns.

18

2.4 The importance of a temporal perspective

An explicit temporal perspective in the study of contemporary cities is crucial for

several reasons. First, cities are highly dynamic landscapes and, therefore, a dynamic

approach is required to study them. Indeed, the configuration, composition, and function

of patches in the urban mosaic are dynamic. For example, as urban growth occurs

through infilling, scattered remnant vegetation is cleared in stages, sometimes over

several decades, as different suburbs are developed. Moreover, vegetation condition in

parks and reserves changes as a consequence of natural and human-driven disturbance

regimes and restoration efforts, which vary across time, influenced by climatic and

socio-economic drivers (Luck et al., 2009). Backyard species composition and structure

also change, influenced by gardening practices and fashions, and local socio-economic

drivers (Boone et al., 2009; Luck et al., 2009; Goddard et al., 2010). Furthermore, as

urban populations and demand for land increase, block subdivision and demolition for

higher density construction also increase. Finally, cities can “shrink” because of

population loss, employment decline, and/or economic downturns, which can result in

the passive or forced abandonment of entire neighbourhoods, and commercial and

industrial areas, a phenomenon called ‘de-urbanization’ (Reckien and Martinez-

Fernandez, 2011).

Second, contemporary cities are young and rapidly evolving landscapes (Pyšek et al.,

2010) that have been through recent large-scale habitat destruction and land-use

changes. In such emergent landscapes, remnant ecosystems are likely to be strongly

shaped by past land uses (Foster et al., 2003), and time-lags might mask remnant

biodiversity response to ongoing fragmentation and environmental change (Hahs et al.,

2009; Kuussaari et al., 2009; Jackson and Sax, 2010; Vilà and Ibáñez, 2011).

19

Considering these two aspects is of major importance to the understanding of

biodiversity patterns and processes (e.g., invasion and extinction) in rapidly urbanizing

landscapes.

2.4.1 Land-use legacies

Past land use can affect ecological systems with lasting legacies that persist over time,

sometimes for hundreds to thousands of years (Dambrine et al., 2007; McKey et al.,

2010). These effects can remain even after land-use change and after other more recent

disturbance processes begin operating (Foster et al., 2003; Lewis et al., 2006; Parker et

al., 2010). Depending on the land use within and surrounding remnant ecosystems prior

to urbanization (e.g., agriculture, livestock grazing, or industrial activities), there might

be legacies that influence current biotic and/or abiotic ecosystem components. For

instance, in expanding European cities, the biodiversity in newly formed urban

remnants might have been reduced long before urbanization owing to historical

agricultural land uses (Hahs et al., 2009). Agrarian legacies can also affect urban

remnants soils. In Arizona, for example, residential yards converted from farms had

double the organic matter, nitrogen and phosphorus than yards converted from native

desert (Lewis et al., 2006).

2.4.2 Time-lagged responses to fragmentation

Biodiversity responses to urbanization-induced fragmentation might show a temporal

delay (Kuussaari et al., 2009). This temporal delay depends on several factors, including

the species turnover rate, remnant area, landscape connectivity, and disturbance

intensity. Temporal delays are shorter for species with higher turnover rates (e.g.,

20

annual vs. perennial plants), for smaller and more isolated remnants, and following

small perturbations (Kuussaari et al., 2009). In old cities, remnant biodiversity might be

largely shaped by the disturbance processes originating from the surrounding urban

matrix. However, in young and rapidly urbanizing landscapes, communities are likely to

be gradually adjusting to the novel environment. During this transient period, biological

communities might be better explained by previous rather than current remnant and

landscape spatial configurations (Helm et al., 2006; Kuussaari et al., 2009; Krauss et

al., 2010). Furthermore, these communities contain a transient species pool that might

include species that will go extinct once the transient period is over (i.e., extinction

debt) (Kuussaari et al., 2009). In a similar way, they might not yet be affected by the

invasion of exotic species, which is more likely to occur once remnants re-equilibrate

(i.e., invasion credit) (Jackson and Sax, 2010; Essl et al., 2011; Vilà and Ibáñez, 2011).

Failure to consider land-use legacies effects and transient dynamics in the response of

biodiversity to fragmentation can have major consequences for the scope of urban

ecological research. Indeed, it might lead ecologists to classify remnants with different

fragmentation trajectories and legacies in the same class of urbanization. Essentially,

communities at different stages along the course of adjustment to the surrounding

environment are mixed more or less indiscriminately and independently of their past.

This can lead to incorrect and misleading study design (Figure 2.2) and, ultimately, to

contradictory and unexpected results. Research on the species richness-area relationship

is an example in which misleading interpretations are likely to arise if time is not

considered, because if data were collected in remnants with different ages, any area

effect might be masked by the differences resulting from different trajectories of

fragmentation (Williams et al., 2006).

21

2.5 Towards an emerging framework in urban ecology

A new approach to studying the ecology of cities is needed that incorporates: (i)

awareness that urbanization intensity and age needs to be assessed based on the analysis

of the focal remnant patch and neighbouring landscape, rather then on its position along

a linear geographic transect; (ii) a mechanistic perspective, considering the role of

multiple drivers and their direct and indirect effects on remnant ecosystems; and (iii) an

explicit temporal perspective, acknowledging land-use legacies and time-lagged

responses to environmental change. A more effective emerging framework would

incorporate three essential elements (Figure 2.3).

A first element is a comprehensive set of intervening factors selected using an

ecologically oriented perspective. These factors can quantify drivers, patterns, or

processes, and include: (i) human factors (e.g., socio-economic, demographic, and built

infrastructure); (ii) environmental factors affected by urbanization, including landscape-

scale (e.g., fragmentation) and local-scale factors (e.g., disturbance regimes and local

environmental conditions); and (iii) environmental factors unaffected by urbanization

(e.g., geological and geographical). An ecologically oriented approach is important to

guide the identification of the factors relevant to the ecological question addressed. On

one hand, this requires focus on the species or community of study, because the

response to the environment is species and/or trait specific (Massol et al., 2011;

Schleicher et al., 2011). Therefore, environmental attributes and scales meaningful to

one species or community might not be relevant to another. On the other hand, it

demands a careful analysis of the study area. For instance, if a study is undertaken in a

suburban-type landscape, then income or education level might be more appropriate

drivers of ecological variation in the area than is human population density.

22

Figure 2.3 Dynamic Urban Framework – an emerging framework for the study of cities. The urban-to-

rural gradient approach classifies the degree of urbanization of remnant ecosystems using either

categorical classes or quantitative measures of linear distance between the city centre and the rural matrix

(a) (remnant vegetation in black); or a combination of those with socio-economic, land cover, land use, or

built infrastructure metrics (b) (road density depicted here). Data analysis usually focuses on the

comparison between ecological responses across different urban classes or on the single effects of a

simplified set of explanatory variables (c). A more comprehensive Dynamic Urban Framework uses a

temporal perspective that places the focal urban remnants in their trajectory of change, analyzing the

length of time they have been in the urban landscape and their past spatial configurations and past land

uses (d). It also uses an ecological perspective that identifies the variables that best describe the range of

variation in the community or ecological process of interest (d). A hierarchical perspective is used to understand the causal and interacting relationships between multi-scale drivers and their direct and

indirect effects on the ecological community or ecological process of interest (e).

A second element is an explicit temporal perspective. Given that urban and landscape

ecology have focused mostly on the analysis of spatial patterns, a deliberate shift is

needed from that purely spatially oriented approach towards a perspective that

recognizes landscapes with two main vectors of change: space and time (Gillson, 2009).

From a theoretical perspective, this involves the incorporation of important conceptual

23

constructs, such as extinction debts (Kuussaari et al., 2009; Krauss et al., 2010),

invasion credits (Jackson and Sax, 2010; Vilà and Ibáñez, 2011), and land-use legacies

(Foster et al., 2003). From a methodological perspective, it demands consideration of

the intervening factors mentioned above from a spatial and temporal perspective. Here,

we underline the importance of characterizing the temporal dynamics of landscape

change, including urbanization or remnant age, past remnant and landscape attributes,

and fragmentation drivers (Box 2.3). Temporal dynamics of landscape change are likely

to be particularly important when: (i) urbanization is relatively recent; (ii) there is a

range of time since remnant isolation and/or urbanization; and (iii) there is a range of

previous land uses.

A third element is a conceptual and analytical structure where the relationships between

and among driving and response factors are analysed in a mechanistic manner. This can

be achieved using a hierarchical approach. Urban ecosystems are more likely to be

described as heterarchical rather than hierarchical systems, in the sense that different

factors might or might not be related by causal relationships, depending on the

conditions and scale of analysis (Crumley, 1994, 2007). Nevertheless, a hierarchical

approach provides a middle ground where the complexity of these coupled human-

nature systems can be accommodated, and their multidimensional nature partitioned

into smaller, more manageable sub-systems (Wu and Loucks, 1995; Wu and David,

2002; Qian et al., 2010). The hierarchical patch-dynamic framework (Wu and Loucks,

1995) provides an integrative approach to spatial analysis, whereby the nested structure

of spatial and temporal patterns and processes in urban landscapes can be depicted

(Zipperer et al., 2000). This framework provides core structure to urban LTER projects

in the USA (Grimm et al., 2000; Pickett and Cadenasso, 2006; Pickett et al., 2008) and

its use should be further encouraged. Furthermore, structural equation (Grace et al.,

24

2010) and Bayesian hierarchical modelling (McMahon and Diez, 2007; Qian et al.,

2010) are promising statistical tools to investigate the complex networks of causal and

interacting relationships between multiple factors, and their direct and indirect effects

on remnant biodiversity and ecosystem functioning (McMahon and Diez, 2007; Qian et

al., 2010). Finally, keeping in mind the heterarchical nature of urban ecosystems is

essential because the importance of different ecological-social drivers and their

temporal and spatial boundaries is fluid (Crumley, 2007). This flexibility is fundamental

to a dynamic approach.

Box 2.3 Measuring the temporal dynamics of landscape change

Landscapes are complex systems with two main vectors of dynamism and change: space and

time (Gillson, 2009). Landscape and urban ecology have developed their main body of

knowledge from research on spatial patterns. However, temporal dynamics have often been

ignored and there are very few consistent examples of case studies, nomenclature, or conceptual

frameworks supporting research along the temporal axis of ecological variation. Nonetheless,

historical geographic data, such as aerial photographs and cartographic maps, are available and

can be used to assess how urban landscapes changed through time. Here, it is suggested three

types of variable that can be used to quantify temporal dynamics of landscape change in urban

ecological research.

Urbanization or remnant age – time since the urban patch was developed or the

remnant patch was isolated and surrounded by urbanization, respectively (Loss et al.,

2009; Magle and Crooks, 2009; Park et al., 2010). This variable can be used to quantify

the length of exposure to the urban environment and the time lag in ecological

responses to urbanization-induced fragmentation;

Past remnant and landscape attributes – patch and landscape spatial

configurations (e.g., remnant area and landscape connectivity) (Cousins et al., 2007;

Krauss et al., 2010), land cover (e.g., urban cover), and socio-economic attributes (e.g.,

population density and build-up density) (Boone et al., 2009) can be quantified in a time

series. Future studies could develop time-weighted variables, measuring the age of

attributes of interest;

Landscape fragmentation drivers – variables identifying the main drivers of

landscape fragmentation and isolation of the focal remnant patch (often agriculture,

urban or industrial development). Such variables can be used to track the presence of

land-use legacies.

25

2.6 Application to planning, management, and restoration in

contemporary cities

A critical approach to the assessment of urbanization in ecological studies will expand

the ability and scope of urban ecological research to better intervene in the planning,

management and restoration of remnant ecosystems in contemporary cities. Firstly, a

proper identification of the drivers controlling remnant ecosystems elucidates where

management and restoration efforts should focus, helping to formulate meaningful

management guidelines and tailor strategies of action. This is important not only to

maximize conservation outcomes, but also to minimize costs. Secondly, a temporal

perspective considering land-use legacies and time-lagged ecological responses to

fragmentation places current condition of an ecosystem in the context of its trajectory of

change (Foster et al., 2003), enhancing the understanding not only of observed patterns,

but also the processes and dynamics that generate and maintain them (Box 2.4).

Box 2.4 A temporal perspective in the planning and management of urban remnants

Urban planning

Identification of remnant sizes

Extinction debt research aids in the understanding of how biodiversity varies in time in response

to remnant size and connectivity, variables that are often the scope of urban planning decisions.

Therefore, it can provide guidance on the selection of remnant sizes and landscape

configurations that will allow reasonable conservation outcomes in the future.

Identification and prioritization of remnants to set aside for conservation

A temporal perspective considering the remnant age and past land uses can provide insight into

the biodiversity value of particular remnants and, therefore, can be used in prioritization for

conservation. Priorities could be, for instance, those remnants without significant land-use

legacies and those that were recently fragmented.

26

Management and restoration

Managing and restoring remnants that have land-use legacies

A temporal perspective considering land-use legacies adds realism to the formulation of goals

and understanding of outcomes in restoration. The presence of land-use legacies might mean

that ecosystems have passed biotic and/or abiotic thresholds that might impede restoration

(Foster et al., 2003). Furthermore, if thresholds were crossed, ecosystems are likely to require

specific interventions that are not required in remnants not subject to those legacies. For

example, whereas prescribed burning and mechanical overstorey thinning were important

drivers of the plant community in post-agricultural Pinus palustris woodlands in the south-

eastern USA, these actions had barely any effect on historically forested sites (Brudvig and

Damschen, 2011).

Improving the habitat quality of remnants in transient periods of adaptation

In rapidly urbanizing landscapes where natural areas were cleared for urban development, the

transient period in which remnant biodiversity gradually adjusts to the novel urban scenario

provides a unique opportunity for action (Hahs et al., 2009). Interventions should improve the

habitat quality of these remnants and target: (i) the core patch with restoration efforts and the

design of margins and tracks that minimize influence from humans and external processes; and

(ii) the buffer areas (Paltto et al., 2006), by improving connectivity and enhancing the urban

matrix at various scales, from the individual garden to the neighbourhood or suburb (Goddard et

al., 2010). These interventions should target priority remnants, and also those where keystone

species are present and whose extinction are predicted to have cascade effects on the survival of

other species (Krauss et al., 2010).

2.7 Concluding remarks

In the context of a rapidly urbanizing world, it is important to consider the complex

growth, relative youth and dynamic nature of contemporary cities if ecologists want to

move forward in the study and conservation of the places where most humans live and

work (Miller and Hobbs, 2002). Failure to consider these characteristics compromises

the scope of urban ecological research, potentially leading to ill-designed studies and

partial or misleading research outcomes. Furthermore, it limits the ability of urban

ecology to provide meaningful guidance to planning, conservation, and restoration in

27

cities. Here, it was suggested the essential elements of an emerging Dynamic Urban

Framework. From a conceptual perspective, this framework is based on ecological

theory that urgently needs to be incorporated into mainstream urban ecological research.

In particular, the transient dynamics in biodiversity response to environmental change,

including extinction debts (Kuussaari et al., 2009; Krauss et al., 2010) and invasion

credits (Jackson and Sax, 2010; Vilà and Ibáñez, 2011), implications of land-use

legacies for conservation (Foster et al., 2003), hierarchical patch dynamics (Wu and

Loucks, 1995; Qian et al., 2010), and hierarchical modelling (McMahon and Diez,

2007; Qian et al., 2010), all need to be incorporated. From a practical perspective, the

Dynamic Urban Framework: (i) is grounded in the area and community of study; (ii)

places the process of urbanization and its effects on biodiversity and ecosystem

functioning in a temporal context; and (iii) depicts the observed ecological responses as

the result of multiple measurable factors that relate and interact at different spatial and

temporal scales. As a whole, the conceptual and practical elements of the framework

can be a first step towards the foundation of a new approach to the study of cities.

29

3. DELAYED EFFECTS OF FRAGMENTATION ON PLANT SPECIES

RICHNESS AND ABUNDANCE IN REMNANT WOODLANDS OF A

RAPIDLY URBANIZING BIODIVERSITY HOTSPOT

Contemporary urbanization is leading to the rapid and extensive fragmentation of

natural vegetation into small and scattered urban remnants. Understanding what

factors alter plant communities in these newly formed remnants requires the

consideration of the complex effects of fragmentation in urban areas and how these

might be delayed in time. This study investigated the effects of fragmentation on plant

species richness and abundance in 30 remnant Banksia woodlands of the rapidly

urbanizing city of Perth, located in the south-western Australian global biodiversity

hotspot. It was considered a comprehensive set of factors characterizing landscape

fragmentation dynamics (current and past remnant area and connectivity, time since

isolation, and trajectories of landscape change), disturbance regimes (fire frequency,

grazing intensity, and human activities), and local environmental conditions (soil

nutrient status and litter depth). Structural equation modelling was used to disentangle

the direct and indirect effects of landscape and local factors on plant species richness

and abundance. It was observed that richness and abundance of woody species was

higher in historically large remnants and lower in the more connected rural city

fringes. Richness of native herbaceous species declined with time since isolation, mainly

in the smaller remnants, and with soil organic carbon. An increased abundance of non-

native herbaceous species and litter depth (presumably reflecting higher productivity),

largely in the smaller remnants, drove the decline in the abundance of native

herbaceous species. Richness of non-native herbaceous species was higher in the less

30

connected remnants closer to the city centre. Abundance of non-native herbaceous

species was strongly influenced by herbivory. The study shows that fragmentation in

urban areas has profound and complex effects on remnant plant communities, which

are accentuated by the human-dominated nature of the surrounding matrix.

Importantly, the study shows that in cities with a recent development history and where

urbanization is an important fragmentation driver, impacts of on-going urban

expansion on remnant vegetation might take many decades to manifest themselves.

Nevertheless, the smaller and older remnants already indicate several environmental

and ecological changes due to fragmentation, and should provide clues for urban

planning and management that help diminish the impacts of urbanization on remnant

ecosystems.

3.1 Introduction

The world is experiencing an unprecedented urban transition (UNFPA, 2007; Seto et

al., 2010). Between 2000 and 2030, the world urban population is expected to grow

from three to five billion, and middle-sized cities (500,000 to 1 million people), in

which most of the current urban expansion is occurring, are expected to triple in their

area (UNFPA, 2007). Furthermore, the shape of the cities is also changing, as they are

no longer compact, but increasingly dispersed and expansive (Seto et al., 2010). While

in many cities, urban expansion is taking place largely on agricultural land (Seto et al.,

2000), in others (e.g., west coast of the USA) it is also leading to the rapid and extensive

loss of native vegetation and its fragmentation into small and isolated urban remnants

(Hahs et al., 2009; Seto et al., 2011). As a result, urbanization is a primary cause of

species endangerment (Czech et al., 2000; Burgman et al., 2007) and a threat to the

31

conservation value of protected areas in and near metropolitan regions (Wittemyer et

al., 2008; Radeloff et al., 2010).

The effects of fragmentation in urban areas are complex and diverse and affect remnant

plant communities via a number of inter-related pathways. Fragmentation affects

remnant plant communities via its impacts on colonization-extinction dynamics (sensu

MacArthur and Wilson, 1967). Rare species are mostly vulnerable because smaller

remnants support smaller population sizes, but abundant species can also be affected via

fragmentation impacts on seed dispersal and pollination (Tilman et al., 1994; Aguilar et

al., 2006). Fragmentation also leads to alteration of natural disturbance regimes (e.g.,

fire and grazing) and anthropogenic disturbances in the remnants. The latter can be due

to edge effects, which are spill over effects from the surrounding urban matrix that alter

the local environmental conditions (Saunders et al., 1991; Hobbs and Yates, 2003;

Ewers and Didham, 2006).

Effects of fragmentation on remnant plant communities are not immediate. Effects on

the colonization-extinction dynamics take a transient period where plant populations

slowly “relax” towards new species richness to area “equilibria” (Ovaskainen and

Hanski, 2002), with some species drifting to extinction while others are invading

(Kuussaari et al., 2009; Vilà and Ibáñez, 2011). The length of this transient period or

time-lag depends on several factors, including the species lifespan, remnant area, and

landscape connectivity, and is smaller for species with shorter lifespan (herbaceous vs.

woody species), and for smaller and more isolated remnants (Helm et al., 2006;

Kuussaari et al., 2009). The effects of fragmentation via impact on the local

environmental conditions and disturbance regimes are also likely to have a cumulative

effect over time. Due to the overriding influence of human activities in urban areas,

32

local factors might have a more rapid and stronger impact than the sole effects of

fragmentation on the colonization-extinction dynamics (Williams et al., 2006). A “time

effect” in the older and smaller remnants is likely to encapsulate the multiple effects of

fragmentation on colonization-extinction dynamics and local disturbances (Bolger et al.,

2000).

Current understanding of the effects of fragmentation in rapidly urbanizing landscapes

has been hampered for two main reasons. First, previous studies have mainly analysed

the influence of landscape and local factors individually (e.g., Hamberg et al., 2008), or

more often, using aggregated urbanization measures (e.g., population density, distance

to the Central Business District or CBD) that are assumed to encapsulate the effects of

fragmentation, age and degree of urban development (McDonnell and Hahs, 2008).

Second, despite an increasing recognition of the importance of a temporal perspective

(Lewis et al., 2006; Boone et al., 2009; Hahs et al., 2009), previous studies have

predominantly used a static approach that does not consider the temporal dynamics of

landscape change, and the effect of time-lags and land-use legacies on current

vegetation patterns (Chapter 2). An explicit consideration of the multiple landscape

fragmentation effects and their temporal dynamics is, however, crucial for better

understanding the effects of contemporary urbanization on remnant biodiversity, as well

as allowing more targeted remnant conservation and management actions (Shochat et

al., 2006; Didham et al., 2007; Schlesinger et al., 2008; Chapter 2).

This study aimed to understand the effects of fragmentation on plant species richness

and abundance of 30 remnant Banksia woodlands in the rapidly urbanizing city of Perth,

in the south-western Australian global biodiversity hotspot. To do so, the conceptual

model presented in Figure 3.1 was followed. It was hypothesised that: (a) species

33

richness and abundance might be better explained by past rather than current remnant

area; (b) with time since isolation, richness and abundance of native species might

decrease, whereas richness and abundance of non-native species might increase; these

changes are likely to be faster in smaller than larger remnants (Helm et al., 2006;

Kuussaari et al., 2009); (c) urban remnants originally fragmented for agricultural

development might carry an agrarian legacy in the soil nutrient status (Foster et al.,

2003; Lewis et al., 2006; Chapter 2); (d) which might lead to an increased abundance of

non-native species (Vilà and Ibáñez, 2011); (e) increased fire frequency might also lead

to an increased abundance of non-native species (Fisher et al., 2009a); (f) increased

abundance of non-native species leads to native species displacement.

Non-native

plant species

Native plant

species

Landscape factors

Local factors

Plant community

Fire

frequency

Grazing

intensity

Local env.

conditions

Human

activities

intensity

(a) (b) (b) (c)

(e)

(f)

Past

remnant area

Current

remnant area

Time since

isolation

Landscape

trajectory:

rural-urban

Landscape

trajectory:

rural

(d)

x

Figure 3.1 Conceptual model used in the study. The boxes represent major conceptual constructs. The

cross sign between current remnant area and time since isolation represents the interaction term between

these variables. Thick arrows represent the general relationships that were examined. Narrow arrows represent specific relationships that were hypothesised a priori.

34

3.2 Methods

3.2.1 Study area

The study was conducted in remnant Banksia woodlands of the Perth Metropolitan

Area, Western Australia (31º57’18.64”S, 115º51’30.37”E) (Appendix A). The study

area has a flat to gently undulating topography and is situated on two ancient dune

systems, Spearwood (mid- to late Pleistocene) and Bassendean (early Pleistocene)

(Kendrick et al., 1991). These are composed of well-drained and highly leached pale

yellow (Spearwood) or white (Bassendean) quartz sands, with poorly drained organic

matter-rich soils surrounding the lower interdunal swamps and lakes (Semeniuk and

Glassford, 1989). These coarse-textured soils (97% sand) are extremely poor in

nutrients, have low cation exchange capacity, and are acidic, typically in the pH range

of 4.5–6.0 (McArthur et al., 2004). The climate is Mediterranean, with a mean annual

precipitation of 740 mm, 80% of which falls in the winter months (between May and

August) and only 4% during the summer months (December to February) (Australian

Bureau of Metereology, 2011). The study area is located in the south-west Australian

Floristic Region, which is a global biodiversity hotspot (Myers et al., 2000) with

exceptionally high levels of floristic diversity, endemism, and species geographic

turnover (Hopper and Gioia, 2004).

Banksia woodlands have an open canopy dominated by Banksia attenuata and B.

menziesii and other less-abundant species, including Eucalyptus marginata and

Allocasuarina fraseriana. The species-rich understorey is dominated by sclerophyllous

shrubs from the families Proteaceae, Myrtaceae, Fabaceae, and Epacridaceae, with a

perennial herbaceous community mainly from the families Restionaceae, Cyperaceae,

35

Haemodoraceae, and Dasypogonaceae. Most non-native species are herbaceous and

include members from the Poaceae, Asteraceae, and Iridaceae, and originate from the

Mediterranean basin and Cape Region in South Africa (Dodd and Griffin, 1989).

Perth is one of the most sprawling cities in the world and one of the fastest-growing in

Australia (Australian Bureau of Statistics, 2011), currently extending over 120 km along

the coast and covering 100,000 ha (Weller, 2009). This suburbia-style city is dominated

by 1–2 storey houses at an average density of six houses per hectare (Weller, 2009), and

with only a few, localized industrial and high-density building areas. Perth’s current

population of 1,659,000 is estimated to reach 3 million by 2050 (Australian Bureau of

Statistics, 2011). Clearing and fragmentation of Banksia woodlands started shortly after

European settlement in 1829. Because the sandy soils proved to be inadequate for

agriculture, urbanization has been the main fragmentation driver of these woodlands.

Urban growth peaked from the 1960’s onwards, fuelled by a mining boom that has since

then largely driven the state’s economy (Weller, 2009). As the city expanded, the extent

and configuration of remnant vegetation changed progressively as surrounding suburbs

were developed in stages. Still, the urban matrix remained fairly unchanged, with the

only main vector of landscape transformation being block subdivision.

In the Perth Metropolitan Area, remnant Banksia woodlands persist in a few large

conservation and Crown Land areas on the current city boundaries, and in urban

reserves (most of which are small and isolated), scattered linear strips on roadside

verges, and rural private properties. The extensive and rapid urban sprawl has had large

impacts on these communities through habitat loss, introduction of non-native species

and plant diseases (e.g., Phytophthora cinnamomi), as well as disruption of fire and

grazing regimes (Stenhouse, 2004, 2005; Burgman et al., 2007; Fisher et al., 2009a).

36

Ground water consumption for public metropolitan supply is also known to affect these

plant communities, as decreased water table levels impedes access to water for

phreatophytic species, including all canopy species (Groom et al., 2000).

3.2.2 Remnant selection and sampling design

Geographic data from the Perth Metropolitan Area was assembled from different

governmental agencies in an ArcGis 9.3 database (ESRI, 2008) (Appendix B). Remnant

selection was based on a vectorial remnant vegetation map provided by the Western

Australian Department of Planning and Infrastructure. Remnants were classified as

suitable for sampling if they had an area >1 ha and narrowest width >70 m, were

isolated by other land-use types or by a main road, and were located within a radius of

30 km from the city centre, in Basseandean or Spearwood soils (excluding wet and

periodically waterlogged soils). Remnants selected were then classified into three size

classes (1–5 ha; 5–50 ha; >50 ha) and four classes of time since urbanization (urbanized

by 1965, 1985, 2006, currently still in a rural matrix). These classes of time since

urbanization were chosen in order to track the main landscape changes since the 60's,

when urban expansion in Perth started accelerating. Three remnants belonging to each

size and time since urbanization classes were randomly selected. From the potential

sample of 36 remnants only 30 were selected since there were fewer than three remnants

available in some classes. In particular, no remnants isolated by urbanization in 1965

and with an area >50 ha were available. Final selection was refined and, whenever

possible, remnants with the following characteristics were excluded: a) less than 500 m

apart (to minimize spatial autocorrelation); b) with >30% of the area composed of wet

or periodically waterlogged soils; c) composed of two or more patches divided by minor

37

roads or fences, when different types of management and/or history were evident

between them (e.g., clearing or grazing).

Three, five and seven plots were randomly located in each remnant with 1–5 ha, 5–50

ha, and >50 ha of area, respectively (n = 130). Plots were composed of nested circular

areas in which different community and environmental variables were estimated. These

included two concentric circles of 11 m and 5.5 m radius, plus three circles of 1.5 m

radius located in the north, south and randomly selected east or west 5.5 m circle

locations (configuration of the sampling plot in Appendix C). All plots were located at

least 25 m from the remnant margin, 5 m from main tracks, and had not been burnt in

the previous 5 years.

3.2.3 Vegetation survey

Vegetation sampling took place in the spring and early summer of 2008 and 2009 (peak

flowering season). Within each plot, all vascular plant species were recorded and their

relative abundances visually estimated using 12 classes of cover (+; <1%; 1-4%; 5-9%;

10-14%; 15-19%; 20-24%; 25-34%; 35-49%; 50-74%, 75-89%, 90-100%). Woody and

herbaceous species were recorded in the 5.5 m and 1.5 m radius circles, respectively.

The size of the concentric circles and the cover classes used were determined based on

standard vegetation sampling procedures (Brower et al., 1997) adjusted to the

characteristics of the studied plant community. Mean percent cover per species per plot

was calculated by taking the mean of each cover class. In the case of herbaceous

species, data obtained in the three 1.5 m radius circles were further averaged for the plot

level.

38

3.2.4 Landscape and local factors

Landscape fragmentation dynamics. Remnant area and landscape connectivity were

estimated in two years, 2006 and 1965. Recent landscape configurations (2006) were

obtained using a vectorial remnant vegetation map derived from photo-interpretation

and provided by the Western Australian Department of Planning and Infrastructure,

whereas historical landscape configurations (1965) were obtained by photo-

interpretation of historical aerial photographs. Although seed dispersal distance in

Banksia woodlands is very short, with most species having no specialized seed dispersal

mechanism (Hopper, 2009), landscape connectivity was measured, as this could have an

indirect effect through impact on pollinators and herbivores. Landscape connectivity

was estimated as the percent cover of remnant vegetation in a 2 km radius buffer

centred on each remnant centroid. Other related measures of connectivity commonly

used in the literature (Moilanen and Nieminen, 2002), such as those weighting remnant

area with distance to the next remnant (e.g., Hanski, 1998), could not be used here

because in 1965 a significant part of the landscape was still continuous. Because

remnants were often fragmented in stages, as urban growth occurred through infilling,

remnants were considered isolated when they were surrounded by a different land-use

type and had area smaller than twice the current one. Time since remnant isolation and

time since urbanization (in years) were determined through observation of aerial photos,

Landsat imagery, and historical maps (Appendix B). Remnants isolated for 45 years or

more were given an age of 45 years, given the scarcity of older records available. In this

study, time since isolation was used so the effects of fragmentation on rural remnants

could also be understood. Finally, depending on the main landscape fragmentation

driver (i.e., agriculture or urbanization), remnants were classified in three classes of

landscape trajectory: 1) urban remnants historically fragmented for agricultural

39

development (rural – urban), 2) urban remnants directly fragmented for urban

development (woodland – urban), and 3) remnants still located in a rural matrix (rural).

Local environmental conditions. Twenty surface soil samples (0-10 cm depth) were

randomly collected in the 5.5 m radius sampling circle, using a 3.5 cm diameter auger,

and bulked together for standard soil analysis (Colwell P, Extractable S, Colwell K,

Walkley Black percent organic C, NO3, NH4, and pH (H2O)). Percent cover of bare

ground and litter were estimated in the same area. In addition, litter layer depth (cm)

was measured in the centre and north, south, east, and west positions of the three 1.5 m

radius sampling circles and averaged at the plot level.

Disturbance regimes. Fire frequency and number of years since last fire in the past 30

years was determined for each plot. The occurrence of wildfires, arson, and controlled

burns, was determined through observation of aerial photos and satellite imagery

(Appendix B), as well as records from land management agencies. Grazing intensity by

the native western grey kangaroo (Macropus fuliginosus) and the non-native European

rabbit (Oryctolagus cuniculus), the two herbivores with the greatest impact on Banksia

woodlands, was assessed within the 11 m radius sampling plot through a qualitative

assessment based on Stenhouse (2005). Intensity of human activities, including

trampling, waste disposal, and soil physical disturbance was estimated in similar

manner. Semi-quantitative variables (Appendix D) were scored from 0 to 5 in ascending

order of significance in the site (0 = absent, non-significant; 1 = very low; 2 = low but

significant; 3 = intermediate; 4 = high; 5 = very high). Six final composite variables,

measuring the intensity of trampling, waste disposal, soil physical disturbance, overall

human activities influence, grazing by native and non-native herbivores, were calculated

by summing the different scores and dividing by the total maximum value.

40

Other environmental factors. Distance to the city centre, depth to the water table and

main soil type, i.e., Basseandean or Spearwood, were recorded for each plot using

digital geographic data (Appendix B). Landscape and local factors analysed in the study

are presented in Table 3.1.

Landscape and local factors Units;

transformation Range (median)

Landscape fragmentation dynamics

Current remnant area* ha; log (x+1) 0.7 – 264.4 (17.4)

Past remnant area* ha; log (x+1) 6.5 – 14533.1 (11243.0)

Current remnant connectivity* ha 9.7 – 630.7 (185.0)

Past remnant connectivity* ha 53.5 – 4647.1 (1595.3)

Time since remnant isolation* no. of years 0 – 45 (25)

Time since remnant urbanization* no. of years 0 – 45 (20)

Trajectory of landscape change (rural-urban;

woodland-urban; rural) Unitless (0/1)

Disturbance regimes

Fire frequency+ no. of fires 0 – 6 (2)

Years since last fire no. of years 10 – 40 (18)

Grazing intensity by native herbivores Composite index 0 – 1 (0)

Grazing intensity by non-native herbivores Composite index 0 – 0.65 (0)

Intensity of trampling Composite index 0 – 0.35 (0.05)

Intensity of waste disposal Composite index 0 – 0.55 (0.025)

Intensity of soil physical disturbance Composite index 0 – 0.53 (0)

Local environmental conditions

P (Colwell P) mg Kg-1; log (x+1) 1 – 8 (1)

S (Extractable S) mg Kg-1; log (x+1) 1 – 4.5 (1.7)

K (Colwell K) mg Kg-1; log (x+1) 11 – 125 (25)

NO3 mg Kg-1; log (x+1) 0.5 – 6 (1)

NH4 mg Kg-1; log (x+1) 1 – 9 (2)

Percent organic C (Walkley Black) mg Kg-1; log (x+1) 0.4 – 3.8 (1.2)

pH (H2O) unitless 5 – 7 (6)

Conductivity dS m-1; log (x+1) 0.009 – 0.06 (0.02)

Cover by bare ground % 0 – 9 (5)

Cover by litter % 3 – 13 (9.5)

Litter depth cm 0.09 – 4.6 (0.8)

Other factors

Depth to the water table (historical minimum) m 3.8 – 53 (15)

Main soil type (Spearwood, Basseandean) Unitless (0/1)

Distance to the CBD Continuous (Km) 4.5 – 32.9 (13.8)

Table 3.1 Description of the landscape and local factors used in the study *+ Variables collected using GIS: *at the remnant level, +at the plot level.

41

3.2.5 Data analysis

A Principal Component Analyses (PCA) of the landscape and local factors was initially

conducted for visualization of the relationships between those factors (Appendix E). All

other statistical analyses were performed separately for woody and non-woody native

and non-native herbaceous species, given the different life history and origin,

respectively, of these groups.

Generalized linear mixed-effect models (GLMM) with Poisson error distribution and

log-link function (Zuur et al., 2009) were used to, first, test the interactive effect

between current remnant area and time since isolation and, second, identify the

landscape and local factors that better explained plant species richness. In these models,

remnant was used as a random effect and landscape and local factors used as fixed

effects. In the models testing the interactive effect between remnant area and time since

isolation, only these two fixed effects and their interaction term were included.

Predictors were centred on their means so that coefficients could be interpreted as the

amount of change in the response variable following a unit change in the predictor,

holding other predictors constant at their mean values (Aiken and West, 1991).

Collinear variables with Pearson correlation coefficient >0.65 were not introduced in

the same model and their effects tested separately. Model selection was based on

Akaike’s Information Criterion (AIC). Residuals were visually inspected to check for

model assumptions (Zuur et al., 2009). Models were constructed using the lme4

package (Bates et al., 2011) in the R environment (R Development Core Team, 2011).

Structural equation modelling (SEM) (Grace, 2006) was used to quantify the

relationships between landscape and local factors and determine their direct and indirect

42

effects on plant species richness and abundance. Thus, SEM models contained a broader

set of environmental factors than those selected in the GLMM models and their results

were the main discussion focus. The conceptual model used to guide SEM analyses is

presented in Figure 3.1. Selection of variables to represent the conceptual entities of

interest followed the next steps. First, among the three classes of landscape trajectory,

the class rural-urban was selected to account for agrarian land-use legacies in the urban

remnants. The class rural was highly correlated with current landscape connectivity

(Pearson coef. r=0.81; Appendix F) and the latter was used instead given its selection in

the best-fitting GLMM models. Second, for local environmental conditions, percent

organic C and litter depth were selected given their selection in the best-fitting GLMM

model for native species richness. Although soil P concentration did not show any effect

on species richness, this variable was selected given its potential effect on the

abundance of non-native species. Best-fitting models were determined by the Chi-

square statistic (a ratio less than two between the chi-square and the number of degrees

of freedom indicates good fit) and p-value (a significant p-value indicates poor fit)

(Grace, 2006). SEM were conducted using Amos software (Arbuckle, 2010).

3.3 Results

A total of 292 plant species, including 115 woody (104 shrubs and 11 trees) and 177

herbaceous species (137 native and 40 non-native), were found across the 30 remnants.

There was high species geographic turnover, with around 70% of the native species only

found in less than 20% of the plots.

43

3.3.1 Interactive effect of time since isolation and remnant area on species

richness

Richness of woody species declined with time since isolation but only in the smaller

remnants (Table 3.2, Figure 3.2). Richness of native herbaceous species also declined

with time since isolation. Although this decline was stronger in the smaller remnants, it

was also significant at average remnant area (Table 3.2, Figure 3.2). Finally, richness of

non-native herbaceous species increased over time in the smaller remnants but

decreased in larger remnants (Table 3.2, Figure 3.2).

Species richness Current remnant

area

Time since

isolation

Current remnant

area x time since

isolation

Woody species ns ns 0.017 **

Native herbaceous species ns -0.006 * 0.013 *

Non-native herbaceous species -0.269 * ns -0.025 *

Table 3.2 Results of the Poisson generalized linear mixed-effect models testing the interactive effect of

current remnant area and time since isolation on the woody and non-woody native and non-native

herbaceous species richness. Standardized β and p values are provided (ns – non-significant; * – p<0.05;

** – p<0.01).

3.3.2 Effects of landscape and local factors on species richness

Richness of woody species was higher in historically large remnants and lower in the

more connected rural city fringes (Table 3.3, Figure 3.3). Richness of native herbaceous

species was similarly influenced by those landscape factors. Furthermore, it was lower

in the remnants with higher litter depth, percent organic C, and soil physical disturbance

(Table 3.3). Richness of non-native herbaceous increased over time in the smaller

remnants but decreased in the larger remnants. Moreover, it was higher in less acidic

soils.

44

Figure 3.2 Interactive effect between time since isolation and current remnant area on the richness of

woody and non-woody native and non-native herbaceous species. The dots represent sampling sites (or

remnants) and the red lines are best-fit lines. In the time span of the study (45 years), richness of woody

(r=0.68; p=0.016) and native herbaceous species (r=0.66; p=0.02) significantly decreased with time since

isolation in the small remnants (1-5 ha), but did not vary significantly in remnants with intermediate (5-50

ha) and large (>50 ha) areas. Non-native herbaceous species richness significantly decreased with time

since isolation in the larger remnants (r=0.82; p=0.02).

45

Woody species

Stand. β SE z p

(Intercept) 2.965 0.032 92.010 <0.001

Past remnant area 0.173 0.035 4.950 <0.001

Current landscape connectivity -0.001 0.000 -4.29 <0.001

Native herbaceous species

Stand. β SE z p

(Intercept) 3.299 0.175 18.873 <0.001

Past remnant area 0.120 0.036 3.353 0.0008

Current landscape connectivity -0.001 0.000 -2.661 0.0078

Soil physical disturbance -0.531 0.229 -2.324 0.0201

% organic C -0.720 0.288 -2.504 0.0123

Litter depth -0.070 0.033 -2.165 0.0304

Non-native herbaceous species

Stand. β SE z p

(Intercept) 1.664 0.075 22.241 <0.001

Current remnant area -0.357 0.125 -2.851 0.0043

Time since isolation 0.002 0.005 0.313 0.7545

Current remnant area x time since isolation -0.026 0.010 -2.654 0.0079

pH 0.612 0.182 3.356 0.0008

Table 3.3 Results of the Poisson generalized linear mixed-effect models testing the effects of landscape

and local factors on the woody and non-woody native and non-native herbaceous species richness.

Figure 3.3 Relationship between richness of woody species and (a) past connectivity (r=0.62, p<0.001),

(b) current connectivity (r =-0.10, p=0.61), and (c) landscape trajectory (p=0.023; 95% CI). This set of

graphics shows that current woody species richness is better explained by past rather than current

landscape configurations. Furthermore, it shows that the relationship with current landscape connectivity,

which is highly correlated with distance to the CBD (r=0.89) and, therefore, with a general urban-to-rural gradient, is strongly shaped by the response to different landscape trajectories. Richness of woody species

is higher in remnants directly fragmented for urban development (woodland – urban). Landscape

trajectory is represented in the three graphics: rural – urban (red), woodland – urban (green), rural (blue).

46

3.3.3 Disentangling the effect of landscape and local factors on species

richness and abundance

Effects of landscape on local factors

All significant paths between landscape and local factors are presented in the best-fitting

structural equation model for native herbaceous species (Figure 3.4c). The intensity of

human activities and litter depth was higher in the older and smaller remnants. Soil P

concentration was also higher in the older remnants, but only in those with smaller area

and lower connectivity. Grazing intensity by native herbivores was higher in larger

remnants, whereas grazing intensity by non-native herbivores was higher in better-

connected remnants. Fire frequency was lower in smaller remnants, as well as in more

connected rural remnants. Percent organic C increased with litter depth and decreased

with grazing intensity by native herbivores. Fire frequency and intensity of human

activities were positively correlated. No significant effects of past remnant area and

agrarian land-uses legacies on local factors were observed (Figure 3.4c).

Direct effects on plant species richness and abundance

The best-fitting structural equation model for the woody plant community (Figure 3.4a)

explained a substantial proportion of the observed variation in species richness

(R2=0.54) and abundance (R

2=0.91). Abundance of woody species was higher in

historically large remnants and lower in the more connected rural city fringes. Woody

species richness was strongly associated with species abundance, but also increased

with fire frequency (Figure 3.4a). In the best-fitting model for the non-native

herbaceous community (Figure 3.4b), species richness (R2=0.85) was lower in the more

connected rural remnants and was almost entirely determined by species abundance.

Non-native herbaceous species abundance (R2=0.53) increased with time since isolation

47

in the smaller remnants and decreased over time in the larger remnants. Furthermore, it

was strongly affected by herbivory, increasing and decreasing with grazing by non-

native and native herbivores, respectively. In the best-fitting model for the native

herbaceous community (Figure 3.4c), species richness (R2=0.92) decreased with time

since isolation, more strongly so in the smaller remnants (cf. Table 3.2, Figure 3.2).

Furthermore, it was lower in remnants with higher percent organic C (Figure 3.4c).

Past

remnant area

Current

remnant area

Landscape

connectivity

Woody

species

richness

Woody

species

abundance

Fire

frequency

Current

remnant area

Time since

isolation

Landscape

connectivity

Non-native

herbaceous species

richness

Non-native

herbaceous species

abundance

x

Non-native

herbivores

Native

herbivores

Current

remnant area

Time since

isolation

Landscape

connectivity

Native herbaceous

species abundance

Native herbaceous

species richness

Non-native

herbaceous species

abundance

x

Fire

frequency

Native

herbivoresLitter depth

Human

activities

Non-native

herbivoresSoil P

Soil organic

C

(a) (b)

(c)

-0.53 0.94

0.50-0.36

0.13

0.62

-0.16

0.70 -0.30 0.31

0.49 -0.43

0.62 0.50

-0.32-0.39

-0.49 0.43

0.920.94

-0.12

-0.36-0.39

0.11

0.12

0.42

-0.30

0.84-0.72

R2=0.30

R2=0.91R2=0.54

0.31

-0.36

0.70

R2=0.53 R2=0.85

R2=0.39 R2=0.49

R2=0.53 R2=0.63 R2=0.92

R2=0.39 R2=0.49 R2=0.28 R2=0.44R2=0.45R2=0.41R2=0.33

-0.29

0.47

-0.320.32-0.39

-0.51

Figure 3.4 Best fitting structural equation models examining the effects of landscape and local factors on

the richness and abundance of the (a) woody species (df=12; χ2=14.34; p=0.28); (b) non-native

herbaceous species (df=14; χ2=18.26; p=0.19), and (c) native herbaceous species (df=60; χ

2=59.68;

p=0.49). Straight single-headed arrows represent significant effects, whereas curved, double-headed

arrows represent significant correlations. Green and red arrows represent positive and negative effects,

respectively. All the significant paths between landscape and local factors are represented in the model

(c). In the remaining models, only significant effects on the plant community are shown. The strength of each path is given by the arrow thickness, which is proportional to the p value (wider p<0.001;

intermediate p<0.01; narrower p<0.05), and the standardized coefficient (i.e., standardized regression

coefficient).

48

Native herbaceous species abundance (R2=0.63) was, however, the variable that had the

strongest direct effect on native herbaceous species richness. Abundance of native

herbaceous species decreased with increasing abundance of non-native herbaceous

species and, to a lesser extent, grazing intensity by native herbivores and litter depth.

3.4 Discussion

3.4.1 Delayed effects of fragmentation on urban remnant vegetation

The study suggests that in the relatively young and rapidly expanding city of Perth,

remnant Banksia woodlands are only slowly responding to relatively recent large-scale

habitat destruction and landscape change, but that this response is faster in smaller

remnants and dependent on the plant functional group. The smaller and older remnants

(1-5 ha) had approximately just above half of the native woody and non-woody species

of similarly sized remnants that were recently fragmented. This steeper decline in native

species diversity with time in smaller remnants has been described before (Bolger et al.,

2000; Ross et al., 2002) and suggests that smaller remnants are more rapidly and

strongly affected by the multiple fragmentation effects than larger remnants. Those

include the direct effects on the colonization-extinction dynamics and the indirect

effects on the natural disturbance regimes and anthropogenic disturbances (Bolger et al.,

2000). Woody species richness and abundance was, however, primarily explained by

past remnant area, suggesting a delay in the response of the woody plant community to

fragmentation.

49

3.4.2 Effects of landscape connectivity and agrarian land uses

Larger current landscape connectivity was associated with a decline in richness of non-

native herbaceous species and woody species abundance. Current landscape

connectivity was highly correlated with the factors rural and distance to the CBD and,

therefore, with a general urban-to-rural gradient. The higher non-native herbaceous

species richness in the more isolated and older remnants closer to the city centre mirrors

findings from other studies (e.g., McKinney, 2001; Duguay et al., 2007; Gavier-Pizarro

et al., 2010). These remnants have had longer exposure and proximity to the main and

older networks of transport and trade and, therefore, to the main pathways of

introduction, where propagule pressure is higher (McKinney, 2001; Hulme, 2009). The

lower abundance of woody species in the more connected city fringes is counter-

intuitive but may be explained by the city development history and agrarian land uses.

Large sections of rather pristine woodlands in what is today the city’s central belt were

kept undeveloped until around 20 years ago, while further away from the city centre, the

landscape was cleared and shaped by agricultural practices for a longer period. In

particular, in the past, Banksia woodlands within and adjacent to rural properties were

used for livestock grazing. This grazing and trampling disturbance might have reduced

the woody understory cover, leaving a legacy that is still present today, despite the fact

that these practices have ceased for several decades (Bellemare et al., 2002; Flinn and

Vellend, 2005). These results provide an example of the importance of a contextual

approach considering land-use history, in the analysis of the ecological conditions in

urban remnants (Chapter 2).

50

3.4.3 Natural and anthropogenic disturbances

Fire frequency was higher in larger remnants and lower in the rural, less populated

fringes of the city. Despite this known positive association with remnant area (Ross et

al., 2002) and population density (Syphard et al., 2009), fire occurrence in urban

remnants is highly variable and dependent on complex factors, such as socio-economic

context, ownership and management (Guyette et al., 2002; Syphard et al., 2009). In the

study area, for instance, remnants with higher incidence of human activities were burnt

more often. Regarding the impacts of altered fire frequency on the plant community,

fire suppression, mostly in the rural remnants, was associated with a decline in richness

of woody species. This decline is likely to represent obligate seeders, which are not able

to recruit, as adult populations reach senescence and the seed bank becomes inviable

(Regan et al., 2010). Although increased fire frequency is known to promote the spread

of invasive species in remnant Banksia woodlands (e.g., Ehrharta calycina, Fisher et

al., 2009a), such an effect was not observed. This may indicate a heterogeneous fire

history across the urban remnants in terms of other fire attributes, such as fire season,

known to be relevant to invasion dynamics in Banksia woodlands (Hobbs and Atkins,

1990).

Grazing intensity by native and non-native herbivores was influenced by distinct

landscape factors and affected remnant plant communities in different ways. While

grazing intensity by the native western grey kangaroo (Macropus fuliginosus) was

mostly associated with larger remnants where the species still persist, grazing by the

non-native European rabbit (Oryctolagus cuniculus) was higher in the more connected

rural remnants. Native and non-native herbivores had opposing effects on non-native

herbaceous species abundance, a phenomenon that has been described before (Parker et

al., 2006). Indeed, native herbivores had a negative effect on non-native herbaceous

51

species abundance, as they preferably consume non-native rather than native plant

species, a mechanism explained by the biotic resistance hypothesis (Maron and Vilà,

2001; Levine et al., 2004; Morrison and Hay, 2011). In contrast, non-native herbivores

had a positive effect on non-native herbaceous species abundance, perhaps not by

reducing the abundance of native species, as Parker et al. (2006) suggested, but by

disturbing the top 5 cm of soil. Disturbance of this thin layer in the ancient and

impoverished soils of the south-western Australia is known to promote invasion

(Hopper, 2009), as it provides an opportunity for germination of non-native species

seeds abundant in the topsoil, which germinate and grow faster than the native species

(Fisher et al., 2009b).

Disturbance intensity by human activities in the remnants, including recreation and

other non-authorized activities (e.g., waste disposal and motorized sports), was higher in

the older and smaller urban remnants. Although soil physical disturbance, namely by

trampling, is known to negatively affect plant communities by direct damage, alteration

of the micro-environment (Hamberg et al., 2008), and consequent invasion by non-

native species (Hobbs and Huenneke, 1992; Theoharides and Dukes, 2007), these

effects were not detected in the studied Banksia woodlands. This may be because

human activities in the remnants are still largely restricted to edge environments and

tracks (Hamberg et al., 2008), which were deliberately excluded in the sampling design.

Soil P concentration was higher in the older and smaller urban remnants. Although most

of these remnants were initially fragmented by agricultural development, soil data did

not reveal an agrarian legacy, because current rural remnants had lower soil P

concentrations. Instead, results seem to indicate an increased soil fertility associated

with the slow and cumulative effect, over time, of a variety of urbanization-related

52

disturbances, including runoff from surrounding paved and fertilized areas, and

atmospheric pollution (Kaye et al., 2006; Koerner and Klopatek, 2010; Park et al.,

2010). Nevertheless, soil P concentrations were very low, with one third of the remnants

having a mean value of 1 mg Kg-1

or less of Colwell P (representing “readily available”

P). The smaller and older urban remnants had a mean value of 2.75 mg Kg-1

of P. These

low nutrient values might explain why any significant effect of P enrichment on the

plant community was observed, namely on the abundance of non-native species. It

needs to be remembered though that most of the older urban remnants are public parks

with variable weed control management practices, which are likely to have a

confounding effect in the results.

The observed increase in litter depth in the smaller and older urban remnants might be

due to a number of factors, including increased productivity with consequent change in

the community composition and increased litter fall (Grigg et al., 2000) and decreased

decomposition rate (Carreiro et al., 1999). Grigg et al. (2000) found that Banksia

woodland woody species at the edge of a remnant exposed to fertilizer drift from

bordering agricultural land had higher productivity, shoot biomass allocation, and litter

accumulation than individuals in the remnant core. Increased productivity, due to soil

nutrient enrichment might be, therefore, the mechanism explaining the increased litter

depth in the older and smaller urban remnants. Increased litter depth was associated

with a reduction in the abundance of native herbaceous species. Furthermore, increased

percent organic C (possibly due to the higher litter accumulation) was associated with a

reduction in the richness of native herbaceous species. These factors are likely to lead to

higher levels of shade and soil nutrients, which favour germination and growth of non-

native in detriment of native species.

53

3.4.4 Final remarks

The results obtained in the study suggest that future research in remnants of rapidly

urbanizing cities should address two important aspects. First, response of remnant

vegetation to recent fragmentation might be time-lagged and influenced by land-use

history (Hahs et al., 2009). Therefore, consideration of the temporal dynamics of

landscape change, including the time since remnants were isolated, their historical

spatial configurations and land-uses, is fundamental to understanding current and

predicting future remnant vegetation patterns (Chapter 2). Second, given the complex

and inter-related nature of fragmentation effects in urban areas, it is important to

consider the role of individual factors and their causal and interacting relationships

(Shochat et al., 2006; Didham et al., 2007; Schlesinger et al., 2008; Chapter 2). Such an

approach provides an integrative perspective on the factors that directly and indirectly

influence remnant vegetation and that need to be accounted for in their planning and

management. Importantly, it provides information about the effects of remnant area on

the medium and long-term ecological status and conservation capacity of the remnants.

Small remnant Banksia woodlands (1-5 ha) are highly vulnerable to fragmentation and

lose nearly half of their native plant species richness in only a few decades after their

isolation. This decline is likely to encapsulate the effects of fragmentation on

colonization-extinction dynamics but, predominantly, given the steep decline in such a

relatively short period, the indirect effects on the disturbance regimes. Other authors

have reached similar conclusions (e.g., Williams et al., 2006). Smaller and older urban

remnants were more fertile, presumably more productive, and had higher soil organic C

and litter depth. Furthermore, fire frequency and grazing intensity by native herbivores

were predominantly suppressed in the small remnants. These factors are likely to have a

synergistic effect and lead to the rapid decline of native species diversity. While the

54

same patterns were not observed in the larger remnants, the extent to which they will be

affected by those factors is currently unknown and will unfold with time and depend to

some extent on local management activities. Nevertheless, larger remnants are more

resilient and take longer to respond to fragmentation effects (Kuussaari et al., 2009).

Finally, the study area is set in one of the oldest and most infertile landscapes on the

planet, marked by a long evolutionary history with no major climatic (e.g., glaciations)

or deposition (e.g., volcanism) events (Hopper, 2009). The flora of this region is

considered to be resilient to fragmentation given their low seed dispersal and ability to

persist in naturally fragmented landscapes over long evolutionary periods (Hopper,

2009). While this might be true in terms of fragmentation effects on the colonization-

extinction dynamics, the spill over effects from the urban matrix (including soil

disturbance and nutrient enrichment) are likely to reduce significantly the long-term

conservation capacity of small urban remnants in the region. Therefore, decisions on the

size of the remnants to set aside during urban planning are of major importance for plant

conservation in this rapidly urbanizing hotspot. Additionally, while this study was set in

a < 200 years-old city, where urbanization has been one of the main habitat

fragmentation drivers, other cities worldwide are much older and are expanding largely

over agricultural land or (semi-)natural habitats in a “recovery succession” following

historical land uses. The effects of fragmentation on remnant vegetation in expanding

cities with such different development histories can be expected to be very different,

with different extinction debts and extinction rates depending on the cities’

fragmentation history (Hahs et al., 2009). Clearly, comparative studies need to include

both different modes and rates of urbanization, and different biogeographic settings for

a comprehensive picture of the effects of urbanization to emerge.

55

4. PLANT FUNCTIONAL COMPOSITION OF REMNANT

WOODLANDS IN A RAPIDLY URBANIZING BIODIVERSITY

HOTSPOT

Contemporary urbanization is driving the rapid and extensive loss of native vegetation

and its fragmentation into small and isolated urban remnants. These remnants can be in

relatively intact condition yet undergo a slow process of change due to fragmentation-

related effects. In this study, a plant functional trait approach was used to determine the

effects of urbanization-induced fragmentation in 30 remnant Banksia woodlands of the

rapidly expanding city of Perth, in the south-western Australian global biodiversity

hotspot. In particular, the study aimed to understand how the plant community

functional composition changes with time since isolation and what are the key

environmental filters that may underlie those changes. The relative abundances of five

plant functional traits (growth form, seed dispersal, pollination syndrome, nutrient

acquisition, and regeneration strategy) were measured. Fourth-corner analysis and

structural equation modelling were used to examine the influence on the community

functional composition of a comprehensive set of factors characterizing landscape

fragmentation dynamics, natural (fire and grazing) and anthropogenic disturbances,

and local environmental conditions. Overall, most of the changes in the community

functional composition with time since isolation were only observed in the smaller

remnants. In the overstorey of these remnants, an increased relative abundance of trees,

predominantly of ectomycorrhizal species, was observed. Higher nutrient and possibly

water resource availability (e.g., runoff from surrounding paved, irrigated, and

fertilized areas) may explain those patterns. In the understorey of the smaller remnants,

56

the relative abundances of insect-pollinated and animal-dispersed species declined with

time since isolation, possibly due to the effects of fragmentation on their biotic agents.

The relative abundance of wind-dispersed species was higher in smaller and in rural

remnants, likely due to higher soil disturbance through edge effects and grazing,

respectively. This study suggested that an approach using plant functional traits and

considering the landscape fragmentation dynamics might help us understanding the

trajectory of change of remnant plant communities in rapidly expanding cities, and thus

be proactive in the urban planning and remnants management.

4.1 Introduction

In many cities worldwide, urbanization is driving the rapid and extensive loss of native

vegetation and its fragmentation into small and isolated urban remnants (Hahs et al.,

2009; Seto et al., 2011). These urban remnants can be in relatively intact condition

(Hahs et al., 2009) but are likely to undergo a slow process of change in response to

fragmentation effects and anthropogenic disturbances originating from the urban matrix

(Williams et al., 2009). Understanding how plant community composition and

ecosystem function in these newly formed remnants will change over time has major

conservation and social implications. Indeed, such understanding is fundamental to

design guidelines for urban planning and management aiming to prevent the loss of

native biodiversity and invasion of non-native species, as well as maximize the

provision of ecosystem services (Bolund and Hunhammar, 1999; Fuller et al., 2007; Jim

and Chen, 2009) to an increasing human urban population (UNFPA, 2007).

Functional trait-based approaches have been recognized as a valuable framework to

predict changes in plant community composition and associated changes in ecosystem

57

function in response to global change drivers (Lavorel and Garnier, 2002; Chapin, 2003;

McGill et al., 2006; Webb et al., 2010). These approaches view plant communities as

the result of a hierarchical set of environmental filters that constrain which species and

traits, from a regionally available pool, can persist locally (Díaz et al., 1999; Webb et

al., 2010). Because plant traits are an expression of the plant physiological responses to

the environment (McGill et al., 2006), functional trait-based approaches can thus help

predicting how environmental change affect plant communities (Lavorel and Garnier,

2002; Chapin, 2003; McGill et al., 2006; Webb et al., 2010).

Several recent studies have used trait-based approaches to predict changes in local floras

due to urbanization (Williams et al., 2005; Lososová et al., 2006; Thompson and

McCarthy, 2008; Burton et al., 2009; Knapp et al., 2010; Vallet et al., 2010; Duncan et

al., 2011). However, despite the common underlying assumption that urban floras are

influenced by a set of environmental factors and associated filters (Williams et al.,

2009), these filters are rarely explicitly considered. Instead, aggregated urbanization

measures that are assumed to encapsulate the effects of fragmentation, age, and degree

of urban development are predominantly used (McDonnell and Hahs, 2008). Exploring

the role of individual filters is, however, important in the study of social-ecological

systems, such as cities, where multiple human and ecological factors act and interact in

complex manners (Liu et al., 2007; Chapter 2). Not considering individual factors might

lead ecologists to conclude that particular functional traits (e.g., seed dispersal) are not

affected in the urban settings (e.g., Vallet et al., 2010), whereas they may in fact be

influenced by unmeasured urbanization-related factors (e.g., remnant size).

Furthermore, it might also explain why comparative studies and meta-analyses often

show high variation in trait responses within and across sites, making it difficult to

identify a set of traits that predict plant responses to urbanization (e.g., Duncan et al.,

58

2011). Therefore, further advances require explicit consideration of a more

comprehensive set of urbanization-related environmental factors, as well as the potential

mechanisms through which they filter and shape plant communities (Shochat et al.,

2006; Webb et al., 2010).

Habitat fragmentation affects urban remnant plant communities in two main ways, each

predominantly involving a different set of environmental filters and plant traits. First,

fragmentation has a direct effect on the colonization-extinction dynamics of the plant

species and the biotic agents they rely on for seed dispersal and pollination (Aguilar et

al., 2006; Damschen et al., 2008; Blakely and Didham, 2010). Second, it has an indirect

effect via alteration of the disturbance regimes and local environmental conditions

(Saunders et al., 1991; Hobbs and Yates, 2003; Ewers and Didham, 2006). In this case,

traits related to resource acquisition and use and regeneration are likely to be

predominantly affected (Lavorel and Garnier 2002). For instance, increased soil fertility

might shift competitive hierarchies between species with different nutrient acquisition

strategies (Lambers et al., 2010; Pekin et al., 2011), and disruption of fire and grazing

regimes might shift the relative abundance of seeder vs. resprouter species (Pakeman,

2004; Pausas et al., 2004).

This study examined the effects of fragmentation on the plant functional trait

composition of 30 remnant Banksia woodlands in the rapidly expanding city of Perth, in

the south-western Australian global biodiversity hotspot (Myers et al., 2000; Hopper

and Gioia, 2004). The study focused on five categorical plant functional traits: growth

form, seed dispersal, pollination syndrome, nutrient acquisition, and regeneration

strategy (seeder vs. resprouter). These traits were selected because they have been

shown to determine plant responses to fragmentation in urban and other fragmented

59

landscapes (Duncan and Young, 2000; Henle et al., 2004; Williams et al., 2005; Aguilar

et al., 2006; Knapp et al., 2010; Vallet et al., 2010; Schleicher et al., 2011). The study

examined the influence of a comprehensive set of landscape and local factors on the

relative abundance of the selected plant functional traits. The factors considered were

shown to influence the study community (Chapter 3) and characterize landscape

fragmentation dynamics (current and past remnant area, current landscape connectivity,

time since isolation, time since urbanization, and trajectories of landscape change),

natural (fire frequency and grazing intensity) and anthropogenic (trampling)

disturbances, and local environmental conditions (soil nutrient status and litter depth).

Two main questions were addressed. First, how does the community functional

composition change with time since isolation? In other words, are there trait states being

consistently favoured or disfavoured over time after habitat fragmentation? Second,

what key environmental filters might explain those changes? It was hypothesised that:

(i) lower remnant area and lower landscape connectivity would lead to a decline in plant

species that depend on biotic agents for seed dispersal and pollination (Aguilar et al.,

2006; Damschen et al., 2008); (ii) because the effects of fragmentation on remnant plant

communities can be time-lagged (Helm et al., 2006; Kuussaari et al., 2009; Chapter 3),

past rather than current remnant area would better explain changes in plant functional

composition; (iii) higher soil fertility due to runoff from surrounding paved and irrigated

areas (Kaye et al., 2006; Koerner and Klopatek, 2010; Park et al., 2010) would shift

competitive hierarchies between species with different nutrient acquisition strategies to

the detriment of those best adapted to highly impoverished soils (e.g., root-cluster non-

mycorrhizal species) (Lambers et al., 2006; Lambers et al., 2008); (iv) higher fire

frequency might favour resprouters over obligate seeders (Bell, 2001).

60

4.2 Methodology

4.2.1 Study area and study remnants

A detailed description of the study area and study remnants is given on pages 34-37.

4.2.2 Landscape and local factors

A detailed description of the landscape and local factors analysed in this study is given

on pages 37-40. Additionally, a principal component analysis (PCA) of the soil data

(Appendix F) was conducted. The first PCA axis represented a gradient of soil fertility

(36.4% of variation explained) and was used as such in the data analysis.

4.2.3 Vegetation survey

A detailed description of the vegetation survey is given on page 37.

4.2.4 Plant functional traits

Functional trait data (Table 4.1) were collected for all vascular plant species with an

estimated cover higher than 5% in at least one of the sampling plots (102 from 292

sampled species). In order to include some herbaceous species that were not locally

abundant but widespread in the study area, traits were also measured for an additional

set of 22 species with an estimated cover between 2 and 4% occurring in more than 25%

of the sampling sites (total of 124 species selected for functional trait analysis)

(Appendix G). For data analysis, only traits states (classes within each functional trait)

with average relative abundance >1% were considered (Table 4.1).

61

Functional traits and traits states

No. of

herbaceous

species

(57)

No. of

shrub

species

(57)

No. of tree

species

(10)

% of

species

Average

relative

abundance

Growth form

Shrub - 57 - 46% 52%

Tree - - 10 8% 22%

Herb 34 - - 27% 12%

Sedge and rush 8 - - 6% 9%

Grass 15 - - 12% 5%

Seed dispersal

Unassisted dispersal 42 54 7 77% 85%

Wind dispersal 14 9 8 19% 42%

Internal animal transport 1 6 1 6% 5%

External animal transport - 3 - 2% 4%

Pollination syndrome

Insect pollination 29 51 9 65% 74%

Bird pollination - 13 9 10% 32%

Wind pollination 23 1 1 19% 23%

Self pollination 2 1 - 2% 1%

Undetermined 5 - - 4% 1%

Nutrient acquisition strategy

Arbuscular mycorrhizal 29 41 2 56% 53%

Root cluster* 15 16 3 25% 40%

N-fixing - 15 1 12% 16%

Ectomycorrhizal - 5 5 4% 12%

Ericoid mycorrhizal - 6 - 5% 5%

None 12 1 1 10% 3%

Parasite 1 - 1 1% 1%

Regeneration strategy

Resprouter 20 44 10 52% 87%

Obligate seeder 37 13 - 40% 13%

Table 4.1 Plant functional traits and trait states analysed. The number of herbaceous, shrubs, and tree

species with each functional trait state is given. Species counts for each trait do not necessarily add to the

total species counts (at the top) because species were allowed to have multiple traits states. Trait states are

ordered by their average relative abundance in the plant community. Data analysis only considered the

trait states with average relative abundance >1%.

*Root cluster include proteoid roots (Proteaceae), dauciform (Cyperaceae), and sand-binding

(Restionaceae) roots.

Information on all traits was assembled from several sources, including literature

sources on regional flora, biological databases, expert knowledge, and unpublished field

data collected and observed in the study area. The following sources were used: (i)

growth form: Western Australian Herbarium (1998-2011); (ii) seed dispersal: Royal

62

Botanic Gardens Kew (2008), Dixon (unpublished data); (iii) pollination syndrome:

Keighery (1980), Whelan and Burbidge (1980), Keighery (1982), Holm (1988), Meney

et al. (1990), Bell et al. (1993), Armbruster et al. (1994), Brown et al. (1997), Dixon

(unpublished data); (iv) nutrient acquisition strategy: Lamont (1982), Pate and Beard

(1984), Pate and Bell (1999), Brundrett (2009), Brundrett (personal communication); (v)

regeneration strategy: Hobbs and Atkins (1990), Pate et al. (1990), Pate et al. (1991),

Bell et al. (1993), Meney et al. (1994), Bell et al. (1995), Bell and Pate (1996), Roche et

al. (1997), Lamont and Groom (1998), Pate and Bell (1999), Bell (2001), Enright

(personal communication). In the case of a few, less common, species where trait

information was not available at the species level, information at the genus and family

level was used instead. Multiple categorical classifications were allowed.

4.2.5 Data analysis

The effects of landscape and local factors on the relative abundance of the selected plant

functional traits were assessed using the fourth-corner method (Legendre et al., 1997).

This approach measures the link between a matrix of environmental variables (m

sampling sites x p environmental variables) to a matrix of species traits (k species x n

traits), by way of a matrix of species abundance (m sites x k species abundance values).

The fourth-corner method produces a matrix containing correlation values between

traits and environmental variables (n traits x p environmental variables). The

correlations obtained were tested using 999 permutations and the resulting p-values

were adjusted using Holm’s procedure for multiple testing (Holm, 1979). The

permutation model used (model 1 in Dray and Legendre, 2008) shuffled values for each

species independently. This model assumes that species distributions are primarily

determined by remnant environmental characteristics, independent of the presence of

63

other species or species assemblages (Dray and Legendre, 2008). Fourth-corner analysis

was conducted with all species together for the analysis of the growth form, and then

separately for the overstorey and understorey, given the different life history of these

groups and potentially different relevant environmental factors.

When the fourth-corner analysis results supported the hypotheses stated at the beginning

of the study, structural equation modelling (Grace, 2006) was used to further evaluate

their underlying mechanisms. The relative abundance of each trait was considered as its

community-weighted mean value, which is calculated using the trait value of each

species weighted by the species relative abundance in each site (Garnier et al., 2004;

Lavorel et al., 2008; Laliberté and Legendre, 2010). Full models were constructed in

which the relevant effects among landscape and local factors were derived from the

findings of the Chapter 3, and all direct effects on the relative abundance of the trait

states analysed were included. The most parsimonious reduced models were obtained

using a stepwise specification search based on Akaike’s Information Criteria (AIC).

SEM were conducted using Amos software (Arbuckle, 2010). All other data analyses

were conducted in the R Environment (R Development Core Team, 2011). Community

weighted mean values were calculated using the package FD (Laliberté and Legendre,

2010). Fourth-corner analysis was conducted using the package ade4 (Dray and Dufour,

2007).

4.3 Results

The relative abundance of shrubs declined with time since isolation, but only in the

smaller remnants (1-5 ha) (Figure 4.1, Table 4.2). Furthermore, it was lower in the

remnants that were infrequently burnt, nutrient-enriched, historically small, and that

64

were originally fragmented for agriculture (Table 4.2). The relative abundance of trees

varied in the opposite way, and was higher in the older and smaller remnants (Figures

4.1 and 4.2, Table 4.2). The relative abundance of sedges and rushes increased with

time since isolation, independently of remnant area (Figure 4.1). Additionally, it was

lower in remnants that were historically large and grazed by rabbits. The relative

abundance of grasses was higher in rural remnants (Table 4.2).

Figure 4.1 Scatterplots of the relationship between the relative abundance of each growth form and time

since remnant isolation, for three classes of remnant area (1-5 ha, 5-50 ha, >50 ha). Solid lines represent

significant relationships (p>0.05) and dashed lines represent non-significant relationships.

4.3.1 Functional composition of the overstorey plant community

The relative abundance of ectomycorrhizal and root-cluster species varied in opposite

ways, with ectomycorrhizal species reaching higher relative abundance in the older,

historically small, nutrient-enriched remnants and those with high levels of trampling

(Table 4.2).

65

Landscape fragmentation dynamics

Disturbance regimes and

environmental conditions

Tim

e si

nce

urb

aniz

atio

n

Tim

e si

nce

iso

lati

on

Pas

t re

mnan

t ar

ea

Curr

ent

rem

nan

t ar

ea

Lan

dsc

ape

connec

tivit

y

Tra

ject

ory

rura

l-urb

an

Tra

ject

ory

woodla

nd

-urb

an

Tra

ject

ory

rura

l-ru

ral

Non

-nat

ive h

erbiv

ore

s

Nat

ive

her

biv

ore

s

Fir

e fr

equen

cy

Tra

mpli

ng

Soil

fer

tili

ty

Lit

ter

dep

th

Growth form

Shrub - - + - + + + - - -

Tree + + - + - - + +

Sedge and rush + + - + - -

Herb - -

Grass - + + -

Overstorey


Arbuscular mycorrhizal +

Root cluster - - + + - + - - -

Ectomycorrhizal + + - - + - + + +

N-fixing +

None

Understorey


Insect pollination - - + - -

Wind pollination + - + +

Bird pollination + + + - -

Seed dispersal

Unassisted dispersal + +

Wind dispersal - + - +

Internal animal transport - - + - - -

External animal transport + - + -


Arbuscular mycorrhizal - - + + - +

Root cluster + + - - + - - + + +

N2-fixing + - + - - +

None + - + - + +

Ecto mycorrhizal - +

Ericoid mycorrhizal - - + - - -

Table 4.2 Summary of the fourth-corner analysis results. Significant relationships between each

functional trait state and the landscape and local factors across the 30 remnant Banksia woodland are

shown. Positive relationships are indicated with a positive (+) sign and negative relationships are

indicated with a negative (–) sign. Relationships with p<0.05 are filled in yellow and relationships with

p<0.01 are filled in orange. Only traits where significant correlations were obtained are represented. In

the overstorey, nutrient acquisition strategy was the only trait with significant relationships with the

environmental factors because most species share the same states for the remaining functional traits.

66

When the relative abundance of ectomycorrhizal and root-cluster tree species were

calculated considering the whole plant community, SEM indicated an increase in the

relative abundance of both but predominantly ectomycorrhizal species in the older and

smaller urban remnants (Figure 4.2). The higher relative abundance of ectomycorrhizal

species was associated with higher soil phosphorus concentration. The relative

abundance of root-cluster species was lower in the larger remnants. Both functional

groups but predominantly ectomycorrhizal species were associated with greater litter

depth (Figure 4.2).

-0.32

RuralCurrent

remnant area

Time since

isolationx

Cluster-rooted

species

Ectomycorrhizal

species

Herbivory

intensityPTrampling Litter depth

Fire

frequency

0.36 -0.37 0.54 0.30

0.30

-0.460.26

-0.46

-0.39

-0.28

0.36 0.790.43

-0.32

RuralCurrent

remnant area

Time since

isolationx

RuralCurrent

remnant area

Time since

isolationx

Cluster-rooted

species

Ectomycorrhizal

species

Herbivory


Fire

frequency

Herbivory


Fire

frequency

0.36 -0.37 0.54 0.30

0.30

-0.460.26

-0.46

-0.39

-0.28

0.36 0.790.43

Figure 4.2 Structural equation model of the relative abundance of ectomycorrhizal and root-cluster tree

species (χ2=33.03, df=35, p=0.56). Arrows represent causal pathways from predictor to response

variables, and the number on each path is the standardized partial regression coefficient. Positive

pathways are represented in green and negative pathways are represented in red. The statistical

significance of the regression coefficient is indicated by the width of the arrows. Non-significant

relationships retained on the final model are represented in dashed lines.

4.3.2 Functional composition of the understorey plant community

Pollination syndrome: The relative abundance of insect-pollinated species was lower in

older and smaller remnants, in contrast to wind-pollinated species, which reached higher

abundance in those remnants (Table 4.2, Figure 4.3). The relative abundance of bird-

pollinated species was higher in currently and historically large remnants, and that of

67

insect-pollinated species was higher only in historically large remnants. In contrast, the

relative abundance of wind-pollinated species was lower in both (Table 4.2).

Figure 4.3 Scatterplots of the relationship between the relative abundance of each functional trait state in

the understorey and time since remnant isolation, for three classes of remnant area (1-5 ha, 5-50 ha, >50

ha). Solid lines represent significant relationships (p>0.05) and dashed lines represent non-significant

relationships.

68

Seed dispersal: The relative abundance of species with unassisted seed dispersal was

higher in frequently burnt and larger remnants (Table 4.2). In contrast, the relative

abundance of species with wind-dispersed seeds, which include mostly non-native

grasses, was higher in the smaller remnants. Moreover, it was higher in the rural

remnants and those grazed by rabbits (Table 4.2). This pattern was only observed in the

herbaceous community (Appendix H). The relative abundance of species with seeds

dispersed after animal ingestion decreased with time since isolation in the smaller

remnants (Table 4.2, Figure 4.3). Additionally, it was lower in historically small,

nutrient-enriched remnants and those with high levels of trampling (Table 4.2).

Nutrient acquisition strategy: The relative abundance of root-cluster species increased

with time since isolation, independently of remnant area (Figure 4.3). Furthermore, it

was higher in historically small, nutrient-enriched, frequently burnt, and trampled

remnants (Table 4.2), and lower in those grazed by rabbits and located in the rural

matrix (Table 4.2, Figure 4.3). These patterns were similarly observed for shrubs with

proteoid roots (Proteaceae) and herbaceous with dauciform (Cyperaceae) and sand-

binding (Restionaceae) roots (Appendix H). The relative abundance of ericoid

mycorrhizal species was lower in the older and historically small remnants. The relative

abundance of N2-fixing and species without specialized nutrient acquisition strategy was

higher in the older and nutrient-enriched urban remnants. No consistent significant

effects were observed in the relative abundance of arbuscular mycorrhizal species

(Appendix H).

Regeneration strategy: The relative abundance of obligate seeder and resprouter species

varied in opposite manners in the shrub and herbaceous communities (Appendix H).

The relative abundance of herbaceous obligate seeders decreased with time since

69

isolation, and was lower in remnants with a thicker litter layer and grazed by rabbits. No

significant effect of fire frequency was observed.

4.4 Discussion

The results suggest that the functional composition of the species-rich remnant Banksia

woodlands change over time after fragmentation, in response to several urbanization-

related environmental filters. These changes were more frequently explained by

historical than current remnant area. Furthermore, they were predominantly observed in

the smaller remnants (1-5 ha). These results indicate that the plant functional responses

to recent fragmentation are delayed (Helm et al., 2006; Kuussaari et al., 2009), and that

smaller remnants are more rapidly and strongly affected by fragmentation than larger

remnants (Bolger et al., 2000; Ross et al., 2002; Kuussaari et al., 2009).

Changes in the overstorey

In the smaller urban remnants, it was observed a shift with time since isolation towards

an increased canopy cover and changed canopy composition, with a higher proportion

of ectomycorrhizal species (e.g., Eucalyptus marginata, Allocasuarina fraseriana).

Urban remnants are often exposed to higher nutrients and water availability due to

runoff from surrounding paved, irrigated, and fertilized areas, as well as atmospheric

pollution (Bidwell et al., 2006; Kaye et al., 2006; Koerner and Klopatek, 2010). In

ecosystems with low nutrient and water availability, such as Banksia woodlands, higher

resource availability originating from the urban matrix may lead to changes in the plant

community composition and to higher ecosystem productivity (Martin and Stabler,

2002). Greater abundance of ectomycorrhizal species (e.g., Eucalyptus spp.) over

cluster-root species (e.g., Banksia spp.) in south-western Australian woodlands with

70

greater soil resource availability has been observed before (Lambers et al., 2006;

Lambers et al., 2010). Moreover, a similar canopy change has been observed in a repeat

survey (60-year interval) undertaken in one of the oldest remnants in Perth (Crosti et al.,

2007).

Changes in the understorey

The changes in the overstorey of the small urban remnants are likely to affect the

understorey plant community, and may help to explain the observed decline in shrub

relative abundance. Higher canopy cover reduces the amount of light available to the

understorey. This effect may inhibit the growth of most Banksia woodlands species,

which are adapted to an open canopy and strong sunlight exposure (Pate and Bell,

1999). Higher canopy cover also leads to the accumulation of a thicker litter layer (in

the absence of frequent fire) that might prevent seedling recruitment (Facelli and

Pickett, 1991). Litter of ectomycorrhizal Eucalyptus and Allocasuarina species might

also have allelopathic properties (May and Ash, 1990), which could inhibit shrub

regeneration.

The relative abundance of insect-pollinated species declined with time since isolation,

whereas wind-pollinated species increased. As predicted, lower remnant area may be a

key driver underlying this change in the understorey community. Other studies have

also shown that plant species dependent on insect pollinators for sexual reproduction

decline in smaller urban remnants (Pauw, 2007) due to their susceptibility to the

fragmentation effects on their pollinator communities (Aguilar et al., 2006; Winfree et

al., 2009). Fragmentation limits dispersal of insect pollinators, as smaller remnants are

less readily intercepted and have smaller population sizes that are more prone to

extinction, than large remnants. Thus fragmentation leads to a reduction in pollinator

71

visits and, consequently, to pollen limitation (Aguilar et al., 2006). Insect pollinators

can also be affected by other environmental filters contributing to their decline with

time since isolation. For example, topsoil disturbance and higher litter accumulation due

to the changes in the canopy might destroy the nesting habitats of ground-nesting wasps

and bees (Cane et al., 2006). The results seem to indicate that over time the plant

community composition in the urban remnants tends to change, with an increased

relative abundance of plant species that are not dependent on biotic agents for

pollination, including wind-pollinated species (Aguilar et al., 2008).

The relative abundance of bird-pollinated species was higher in currently and

historically large remnants, but was independent of remnant age. This indicates that

birds might not be as affected by fragmentation as insects, because they have higher

dispersal ability and, therefore, are capable of moving across larger distances between

remnants for foraging (Hostetler and Holling, 2000; Damschen et al., 2008). Also, bird-

pollinated species in the study have several bird pollinator species (Phillips et al., 2010).

The local extinction of habitat specialist pollinators might be buffered by the pollination

services provided by generalist pollinators, which might be capable of using suburbia

(e.g., private gardens) as a complementary or even main habitat and food resource

(Chamberlain et al., 2004; Lim and Sodhi, 2004).

Most Banksia woodland understorey native species may not suffer from seed dispersal

limitation in the urban remnants because they generally lack specialized dispersal

mechanisms and have evolved to persist locally (Hopper, 2009). However, non-native

herbaceous species with wind-dispersed seeds were more abundant than native

herbaceous species in the smaller and in rural remnants. A potential common

denominator to these remnants is higher topsoil disturbance due to edge effects (Harper

72

et al. 2005) and a legacy of livestock grazing coupled with larger rabbit populations

(Chapter 3), respectively. In the study region, topsoil disturbance is known to promote

the germination and growth of non-native species (Hopper, 2009), including

competitive grasses, such as the South-African Ehrharta calycina (Fisher et al., 2009b).

These species have wind-dispersed seeds, germinate earlier in the season, and have

higher growth rates (Fisher et al., 2009b). These characteristics give an advantage to

non-native over native species and predispose invasion in soil-disturbed remnants.

These results are tied to the natural history of south-western Australia and differ from

other studies where native plant species have often higher seed mobility and soil

disturbance is not necessarily associated to plant invasion (e.g., Knapp et al. 2008;

Thompson & McCarthy 2008). Finally, similarly to the insect-pollinated species, the

decline of shrubs with seeds dispersed after animal ingestion in the older and smaller

remnants might be associated to reduced populations of native mammals and reptiles

due to limited immigration and habitat degradation (Damschen et al., 2008).

The changes with time since isolation in the relative abundance of understorey species

with different nutrient acquisition strategies do not indicate an influence of elevated soil

nutrients because root-cluster species, which are best adapted to highly impoverished

soils, reached higher abundance in nutrient-enriched remnants. Hence, the obtained

results are likely to be associated with other plant traits and environmental filters. Here,

a few considerations about two groups are drawn, but further research is required to

understand the observed patterns. First, most root-cluster herbaceous species of the

study area (e.g., Desmocladus, Lyginia, and Alexgeorgea genus) are wind-pollinated,

resprouters, and able to form dense mats (Pate et al., 1991). These traits, coupled with

possible tolerance to shadier conditions, may explain the increased abundance of the

functional group in the older remnants. Second, the decline of ericoid mycorrhizal

73

species in the older and smaller remnants could be explained by two reasons. First, most

species within this group have seeds dispersed by animals, chiefly birds but also

reptiles, ants, and mammals (Keighery, 1996) and hence they may be affected by the

decline of their dispersal vectors due to fragmentation effects (Damschen et al., 2008).

Additionally, ericoid mycorrhizal species rely on dense mycorrhizal networks present in

the 20 cm topsoil (Hutton et al., 1996) and disturbance of this soil layer can reduce

ericoid mycorrhizal species richness and abundance (Hutton et al., 1997).

Contrary to what was predicted, the relative abundance of resprouter and obligate seeder

species do not indicate an effect of fire frequency, but might reflect the functional

response to other environmental filters. For example, for the herbaceous community,

thicker litter layer had a negative effect on the relative abundance of obligate seeder

species, possibly by impeding germination due to excessive shading and by acting as a

physical barrier (Facelli and Pickett, 1991).

Final remarks and future directions

This study has shown that the functional response of Perth remnant woodlands to

fragmentation is time-lagged, with most changes observed along an axis of time since

isolation and more strongly so in the smaller remnants (1-5 ha). Clear response patterns

to fragmentation were not easy to detect because different filters have opposing effects

on the plant community and because different trait responses are likely to have different

relative importance and may even mask or counter-act each other. Nevertheless, based

on a priori hypothesized trait-environment relationships, results seem to suggest that in

the overstorey, increased resource availability in the urban matrix (Martin and Stabler,

2002; Kaye et al., 2006) might lead to an increased canopy cover and changed canopy

composition in the smaller remnants, with a higher proportion of ectomycorrhizal

74

species. These changes in the canopy may affect the understorey plant community and

hence explain the observed decline in the shrubs relative abundance. Effects of

fragmentation on insect pollinators, namely through limited immigration and habitat

deterioration, are likely to underlie a general decline in the relative abundance of

understorey insect-pollinated species with time since isolation. This result raises

conservation concerns, given that most Banksia woodland species are insect pollinated.

Finally, it was suggested that soil physical disturbance due to edge effects and grazing

might explain the higher relative abundance of non-native species with wind-dispersed

seeds in small and in the rural remnants. This result provides a cautionary note for

management, highlighting the importance of reducing the effects of recreation-related

trampling in the urban remnants of this rapidly urbanizing global biodiversity hotspot.

Urban areas worldwide share similar environmental changes due to similar land uses

(e.g., housing, commerce, and transportation) (Williams et al., 2009). Furthermore, they

are marked by the presence of a suite of species that are locally abundant across several

cities worldwide (McKinney, 2006). This has led ecologists to suggest that cities are

centres of biotic homogenization (McKinney, 2006), and has encouraged the

identification of a group of plant functional traits that help predicting the characteristics

of urban floras (Williams et al., 2005; Thompson and McCarthy, 2008; Duncan et al.,

2011). Controlling aspects that determine major variation in the plant functional

responses to urbanization might give future studies greater ability to extract consistent

patterns across multiple cities. First, as Duncan et al., (2011) recently concluded, plant

functional responses to urbanization depend on the fragmentation and development

history of the cities. At one end of the scale, plant communities in cities with a recent

development history and expanding largely into intact native vegetation (e.g., Perth,

Cape Town, Los Angeles) are likely to be strongly affected by landscape fragmentation

75

effects, namely via indirect impact on their seed dispersal and/or pollination agents

(Cane et al., 2006; Pauw, 2007). Currently, as the study indicated, these effects are

likely to be only noticeable in the older and smaller urban remnants. At the other end of

the scale, in cities with a long human occupation (e.g., most European cities), some

plant functions, namely seed dispersal, might even be dependent on human activities

(Thompson & McCarthy 2008).

Second, plant functional responses to urbanization also depend on the environmental

and biogeographic settings. The predominant lack of specialized seed dispersal

mechanisms, large dependence on insect and birds for pollination, and high

vulnerability to soil disturbance are characteristics that shape plant functional responses

to urbanization in the ancient and nutrient-impoverished regions of the globe, as in

south-western Australia (Hopper 2009). However, in the younger, post-glacial, and

more fertile regions of the Northern Hemisphere, these trait states and associated

environmental filters might not be as relevant. The occurrence of rapidly urbanizing

landscapes in cities with different development histories and in different biogeographic

settings provides opportunities for comparative and mechanistic studies of the

functional responses of remnant vegetation to contemporary urbanization.

77

5. DRIVING FACTORS OF TREE MORTALITY AND

REGENERATION IN MEDITERRANEAN URBAN WOODLANDS

The effects of urbanization on canopy species of urban woodlands are still poorly

understood. This study investigated patterns of mortality, density, and regeneration of

the four most common tree species in 30 remnant Banksia woodlands of the rapidly

expanding city of Perth (Western Australia). The study focused on the co-dominants

Banksia attenuata and B. menziesii (Proteaceae), and two other species, Eucalyptus

marginata (Myrtaceae) and Allocasuarina fraseriana (Casuarinaceae). Generalized

mixed models were used to examine the relationship between canopy species variables

(i.e., mortality, density of live trees, and density of recruits) and a set of factors

characterizing landscape fragmentation dynamics (current and past remnant area, time

since isolation, and time since urbanization), natural and anthropogenic disturbances

(fire frequency, years since last fire, grazing intensity, trampling), local environmental

conditions (soil nutrient status and litter depth), topographic position (altitude and

depth to groundwater), and distance to city centre. Time since urbanization and past

remnant area explained the largest amount of variation in the canopy. Mortality of B.

attenuata and B. menziesii were lower in the older and historically small urban

remnants (3%) than in the historically large remnants and those currently located in the

city outskirts (25% and 14%, respectively). In contrast, density of E. marginata was

higher in the older urban remnants. Changes in resource availability (e.g., higher water

and nutrient runoff from surrounding paved, fertilized, and irrigated areas) may

underlie these patterns. Despite the lower mortality of B. attenuata in the older urban

remnants, the density of live trees and saplings was lower in those remnants, possibly

due to higher fire frequency and consequent lower seed production.

78

5.1 Introduction

The effects of urbanization on dominant canopy species of urban woodlands are still

poorly understood. Indeed, most studies aiming to understand the effects of urbanization

on remnant woodlands or forests have focused on the understorey plant community

(e.g., Godefroid and Koedam, 2003; Guirado et al., 2006; Vallet et al., 2010). A main

underlying reason of this focus is that shrub and herbaceous species have shorter life

spans and thus respond more quickly to the environmental changes induced by

urbanization. However, studies in other forest ecosystems have shown that overstorey

tree communities can change rapidly in response to anthropogenic impacts.

Fragmentation, for instance, can induce rapid tree mortality through exposure to edge

effects (e.g., higher wind speed) (Esseen, 1994; Laurance et al., 1998a; Mesquita et al.,

1999).

Because trees can have a strong influence on microclimatic conditions (e.g., shade,

moisture, temperature) and ecosystem functioning (e.g., hydrological and carbon cycles)

(Breshears et al., 2005; Ellison et al., 2005), changes in the canopy of urban woodlands

can have several impacts on these ecosystems. Importantly, they may affect the

structure and composition of the understorey, and thus affect the remnants "capacity" to

sustain their plant biodiversity and wildlife (Ellison et al., 2005). They may also affect

the provision of ecosystem services (e.g., water infiltration, sequestration of air

pollutants, microclimatic amelioration, C storage, and aesthetics), which improve the

urban environment and enhance the wellbeing and quality of living of urban dwellers

(Bolund and Hunhammar, 1999; Jim and Chen, 2009). Hence, it is important to improve

current understanding of the factors influencing the dominant tree species in urban

79

woodlands (Airola and Buchholz, 1984; Fahey, 1998; Lehvävirta and Rita, 2002;

Broshot, 2007, 2011).

This study investigated tree mortality and regeneration in 30 remnant Banksia

woodlands of the Perth Metropolitan Area, Western Australia. In a survey undertaken in

2008-2009, the population mortality, density of adult trees, and density of seedlings and

saplings of four species were quantified. These species included the co-dominants

Banksia attenuata R.Br. (Slender Banksia, Proteaceae) and B. menziesii R.Br.

(Firewood Banksia), and two other frequent species, Eucalyptus marginata Sm. (Jarrah,

Myrtaceae) and Allocasuarina fraseriana (Miq.) L.A.S. Johnson (Sheoak,

Casuarinaceae). Then, the effect of a set of environmental factors was tested with the

aim of identifying which factors are most strongly associated with tree mortality and

regeneration in the study area.

The four studied species are endemic to the south-western Australian global biodiversity

hotspot. Both B. attenuata and B. menziesii are 4-6 m tall, whereas A. fraseriana and E.

marginata can reach up to 15 and 40 m, respectively (Western Australian Herbarium,

1998-2011). These species have different nutrient acquisition strategies, with B.

attenuata and B. menziesii having proteoid roots, which are best adapted to soils highly

impoverished in phosphorus (limiting nutrient), and E. marginata and A. fraseriana

forming associations with ectomycorrhizal fungi (Lambers et al., 2006). All species are

phreatophyte (Dodd and Bell, 1993), extracting groundwater up to 8-9 m deep (Zencich

et al., 2002) in the case of B. attenuata and B. menziesii, and 15 m deep in the case E.

marginata (Dell et al., 1983). Both Banksia species are considered facultative

phreatophytes, being able to rely on soil moisture when groundwater is not accessible

(i.e., summers) (Canham et al., 2009), although this ability is most predominant in

80

plants located in upland areas, where water availability is persistently limiting (Zencich

et al., 2002; Canham et al., 2009). All species resprout after fire (Bell, 2001).

Furthermore, all species except for A. fraseriana, which is wind-pollinated and -

dispersed, are bird- and insect-pollinated and have poor seed dispersal ability (Keighery,

1980; Pate et al., 1990).

It was hypothesised that edge effects have contrasting impacts on tree mortality

depending on remnant age. Small and recently fragmented remnants might have higher

tree mortality due to altered micro-climatic conditions (e.g., increased wind speed)

(Esseen, 1994; Laurance et al., 1998b; Harper et al., 2005). However, water runoff from

surrounding paved and irrigated areas (Kaye et al., 2006; Koerner and Klopatek, 2010)

might lower tree mortality in the older and historically small urban remnants. Given a

degree of dependence on groundwater for all four species, tree mortality might also be

influenced by the sites topographic position and depth to the water table (Zencich et al.,

2002). In terms of regeneration, it was hypothesised that trampling, grazing, and soil

fertility might have a detrimental effect through seedling physical damage (Lehvävirta

and Rita, 2002; Hamberg et al., 2008), seedling predation, and increased competition

with non-native species (Aronson and Handel, 2011), respectively. Litter cover might

provide “safe sites” for seed germination and seedling establishment (Kostel-Hughes et

al., 1998), although thick litter layers might have a detrimental effect on regeneration

(Facelli and Pickett, 1991). Finally, high fire frequency might lower seed production

and contribute to tree mortality (Enright et al., 1998; Fisher et al., 2009a).

81

5.2 Methodology

5.2.1 Study area

A detailed description of the study area is given on pages 34-36.

5.2.2 Remnant selection and sampling design

A detailed description of the remnant selection and sampling design is given on pages

36-37.

5.2.3 Canopy survey

A survey of adult trees took place in the spring and early summer of 2008 and 2009, and

a survey of recruits in 2009 only. All standing trees in the 11 m radius-sampling circle

and with diameter at breast height (DBH, 1.37 m) >10 cm were recorded by species,

and identified if alive or dead. This survey mainly assessed deaths in the last 10 years,

as older trees are likely to have fallen, partly decomposed, or burnt. The number of

recruits <25 cm (seedlings) and >25 cm (saplings) were quantified per species in the

four small 1.5 m radius circles and averaged for each plot.

5.2.4 Environmental factors

The environmental factors considered measured landscape fragmentation dynamics

(current and past remnant area, time since isolation, and time since urbanization),

natural (fire frequency, years since last fire, and grazing intensity) and anthropogenic

(trampling) disturbances, local environmental conditions (soil nutrient status, litter

82

depth, and litter cover), topographic position (altitude and depth to groundwater), and

distance to city centre. A detailed description of these factors is given on pages 37-40.

5.2.5 Data analysis

A redundancy analysis (RDA) (Legendre and Legendre, 1998) was initially constructed

to visually depict the relationships among environmental factors and between these and

the mortality (%), density of live trees (no ha-1

), and density of recruits (no ha-1

) of the

four study species. Forward selection was used to identify a parsimonious set of

explanatory factors in the RDA multivariate environment (Dray et al., 2007). Then,

generalized Gaussian linear mixed-effect models (GLMM) (Zuur et al., 2009) were

conducted for identification of the environmental factors best explaining mortality,

density of live trees, and density of recruits of each species. In these models, remnant

was used as a random effect and environmental factors used as fixed effects. For all

response variables, the interactive effect between current remnant area and time since

isolation/time since urbanization was tested with the aim of understanding if smaller

remnants change more rapidly over time than larger remnants (Ross et al., 2002).

Predictors were centred on their means so that coefficients could be interpreted as the

amount of change in the response variable following a unit change in the predictor,

holding other predictors constant at their mean values (Aiken and West, 1991).

Collinear variables with Pearson correlation >0.65 were not introduced in the same

model and were tested separately. Model selection was based on Akaike’s Information

Criterion (AIC). Residuals were visually inspected to check for model assumptions

(Zuur et al., 2009).

83

Three sampling plots with very high density of B. attenuata recruits (1351, 888, and 625

recruits ha-1

) were eliminated from the data analysis to avoid biased selection of

explanatory factors (median density of B. attenuata recruits where regeneration of this

species was observed was 63 recruits ha-1

). Statistical analysis was conducted in the R

environment (R Development Core Team, 2011). RDA, forward selection in RDA

environment, and GLMM were conducted using the packages vegan (Oksanen et al.,

2011), packfor (Dray et al., 2007), and lme4 (Bates et al., 2011), respectively.

5.3 Results

A main underlying axis of variation in the mortality, density of live trees, and density of

saplings across the study remnants was explained by time since urbanization (Figure

5.1, Figure 5.2). Remnants urbanized for longer had in general also been fragmented for

longer, were historically small, and were nearer the city centre (Figure 5.1).

Density of live B. attenuata (no ha-1

) was higher in the rural and recently urbanized

remnants, where also greater mortality of the species was observed, and lower in the

remnants urbanized for longer and those that were recently burnt (Table 5.1, Figure

5.3). Density of B. menziesii was higher in remnants that were historically small (Table

5.1), although this relationship seemed to be mostly due to two remnants (Figure 5.3).

In contrast to B. attenuata, the density of E. marginata increased with time since

urbanization (Table 5.1, Figures 5.2 and 5.3). None of the factors significantly

explained the density of live A. fraseriana.

Density of B. attenuata seedlings (no ha-1

) was higher in currently large remnants

(Table 5.1), whereas density of saplings increased with distance to the city centre (Table

84

5.1, Figure 5.3). In the second best GLMM model (not shown), density of B. attenuata

saplings was also higher in the historically large remnants. No significant influences of

local factors were observed. None of the factors significantly explained the density of

recruits of B. menziesii, E. marginata, and A. fraseriana.

B. menziesii

E. marginata

A. fraserianaB. attenuata

Grazing

Trampling

P

K

C

Cond

Litter depthLitter cover

Z

Area65Area06

Time since isolation

Time since urbanizationDist. to city

centre B. menziesii

E. marginata

B. attenuataFire

Trampling

S

Litter depth

Litter cover

Area06

Dist. to city centre

Depth water table


Years since last fireKC

NO3pH

RD

A a

xis

2 (

13

.3%

)

RDA axis 1 (50.9%)

RD

A a

xis

2 (

26

.8%

)

RDA axis 1 (44.7%)

B. menziesii

E. marginataB. attenuata

Trampling

Litter depth

Litter cover

Area65

Dist. to city centre

pHTime since isolation

K

NH4R

DA

axis

2 (

24

.8%

)

RDA axis 1 (40.8%)

Time since urbanization

A. fraseriana

ZB. menziesii

E. marginata

B. attenuata

Trampling

P

Dist. to city

centre

SArea65

Z

NH4

RD

A a

xis

2 (

20

.7%

)

RDA axis 1 (45.1%)

A. fraseriana

Time since

urbanization

KLitter cover

Litter depth

a b

c d


Figure 5.1 RDA biplots of (a) mortality (%), (b) density of live trees (no ha-1), (c) density of saplings (no

ha-1), and (d) density of seedlings (no ha-1) of the four study species B. attenuata, B. menziesii, E. marginata, and A. fraseriana, in the 30 remnant Banksia woodlands in the Perth Metropolitan Area

(Western Australia). Only explanatory variables with correlations >0.30 with the RDA axes are shown.

Explanatory variables with correlations <0.40 are presented in grey and those with correlations >0.40 are

presented in black. Underlined explanatory variables were selected through forward selection. Adjusted

R2 of the RDA models containing these selected variables explained 19.9%, 39.7%, and 25.3% of the

variation observed in the mortality, density of live trees, and density of saplings, respectively. Forward

selection did not select any explanatory variables in the RDA model for density of seedlings. No

seedlings of A. fraseriana were found in the study remnants.

85

Figure 5.2 Relationships between mortality (%), density of live trees (no ha-1), and density of saplings (no

ha-1) of Banksia attenuata (red), Banksia menziesii (green), Eucalyptus marginata (grey), and

Allocasuarina fraseriana (orange) with urbanization age (i.e., time since urbanization). Analysis of

variance (ANOVA) revealed that urbanization age had a significant effect on the mortality (p=0.044),

density of live trees (p=0.004), and density of saplings of B. attenuata (p=0.010), as well as on the

mortality of B. menziesii (p=0.056*) and density of live E. marginata trees (p=0.018). Tukey’s HSD post-

hoc tests were used to analyse pairwise differences among classes of urbanization age. Significant

pairwise differences are represented with different letters in the graphics. Post-hoc tests could not distinguish significant pairwise differences between the densities of E. marginata trees in the remnants.

Response variable Stand. β SE z P

Tree mortality (%)

B. attenuata

(Intercept) 0.190 0.023 8.229 0.000

Altitude -0.005 0.002 -3.074 0.003


B. menziesii

(Intercept) 0.107 0.014 7.411 0.000


Density of live trees (no ha-1

)

B. attenuata

(Intercept) 186.439 19.942 9.349 0

Time since urbanization -4.016 1.138 -3.529 0.001

Years since last fire 4.359 1.313 3.321 0.001

B. menziesii

(Intercept) 158.436 17.623 8.990 0.000

Past remnant area -35.470 15.550 -2.281 0.030

E. marginata

(Intercept) 27.011 4.717 5.727 0.000

Time since urbanization 0.903 0.268 3.366 0.002

Density of recruits (no ha-1

)

B. attenuata seedlings (<25 cm)

(Intercept) 8.781 2.417 3.632 0.001

Current remnant area 10.006 3.965 2.523 0.013

B. attenuata saplings (>25 cm)

(Intercept) 54.757 8.000 6.844 0.000

Distance to city centre 4.905 1.179 4.160 0.000

Table 5.1 Results of the best-fitting Gaussian generalized linear mixed-effect models testing the effects of

landscape and local factors on mortality (%), density of live trees (no ha-1), and density of seedlings and

saplings (no ha-1) of the four studied species. Only models where significant relationships were obtained

are presented.

86

Figure 5.3 Scatterplots of landscape and local factors that were significantly related to mortality (%),

density of live trees (no ha-1), and density of recruits (no ha-1) in B. attenuata, B. menziesii, and E.

marginata.

5.4 Discussion

Mortality, density, and regeneration of B. attenuata and to a lesser extent B. menziesii

and E. marginata in the Perth Metropolitan Area were largely explained by time since

urbanization and past remnant area. As these variables are correlated, it is difficult to

ascertain causality. Importantly, time since urbanization is not an effect variable per se,

but a proxy of past effects of a combination of environmental factors that influence the

current status of the remnant ecosystems (Pickett, 1990). Some of the local-scale factors

that were measured and whose effects on remnants ecosystems are known to increase

with time since urbanization (e.g., trampling and soil fertility) (Chapter 3) did not have

a significant influence on the patterns observed. While the influence of these factors

cannot be excluded, the results of this study seem to indicate the influence of other

87

factors that were not measured in the study. In particular, it is hypothesised that changes

in water availability may underlie some of the most relevant patterns observed.

Drought due to decreasing groundwater levels could possibly explain the high mortality

of B. attenuata and B. menziesii in the historically large remnants. Water availability is

often a limiting factor for woody vegetation in Mediterranean-type ecosystems

(Martínez-Vilalta et al., 2002; Jacobsen et al., 2007; Quero et al., 2011). In Perth, an

ongoing drying climate since the 1970’s coupled with groundwater extraction for public

metropolitan supply have been lowering the groundwater levels (Commander and

Hauck, 2005). Significant groundwater declines have been particularly observed in the

northern and southern city outskirts, where the extant of remnant vegetation was much

larger in the 1960’s, and which overly two important aquifers that are under currently

intense exploitation (Commander and Hauck, 2005). B. attenuata and to a lesser extent

B. menziesii are sensitive to drought (Groom et al., 2000; Froend and Drake, 2006;

Fitzpatrick et al., 2008), and lowering groundwater levels have been reported to cause

sudden and regional-scale mortality in these species (Groom et al., 2000; Sommer and

Froend, 2011). Additionally, the ability to survive without groundwater access varies

intra-specifically and is predominant in plants located in upland areas, where water

availability is persistently limiting (Zencich et al., 2002; Canham et al., 2009). Plants

located in lowland areas tend to be groundwater-dependent and thus are more

vulnerable to drought-induced mortality when water tables are lowered (Zencich et al.,

2002; Canham et al., 2009). Although this relationship between mortality and altitude is

known for both Banksia species, in this study it was only detected for B. attenuata,

possibly because of its higher sensitivity to drought.

88

Potentially higher water availability may explain the lower mortality of B. attenuata and

B. menziesii in the historically small remnants. These include the oldest urban remnants

(urbanized by 1965 or longer) and those urbanized by 1985 with 1-5 ha of area. Water

availability due to runoff from surrounding paved and irrigated areas is likely to be

higher in these remnants (Kaye et al., 2006), possibly reducing drought-related tree

mortality. Other studies in cities with arid climates or seasons have also shown

significant effects of water runoff in remnant vegetation. For instance, in the Phoenix

Metropolitan Area (USA), plants in remnants surrounded by irrigated residential areas

had higher water status and productivity than non-irrigated remnants (Martin and

Stabler, 2002). Additionally, municipal water discharges might also raise groundwater

levels (Appleyard, 1995; Kaye et al., 2006).

Other factors known to cause tree mortality are unlikely to explain the observed

mortality patterns of B. attenuata and B. menziesii. Mortality induced by altered micro-

climatic conditions (e.g., increased wind speed) in edge environments of recently

fragmented remnants (Esseen, 1994; Laurance et al., 1998b; Mesquita et al., 1999) was

not observed, as no significant effect of current remnant area was detected. The

pathogen Phytophthora cinnamomi might have caused some mortality in the study

remnants (Shearer and Dillon, 1996). However, the effects of this pathogen are more

localized (Hill et al., 1994) and affect B. attenuata and B. menziesii in a similar manner

(Mccredie et al., 1985), which makes it unlikely to explain the different mortality

patterns observed between the species. Factors such as insect outbreaks (e.g., Simard et

al., 2011), storms (e.g., Comita et al., 2010), and atmospheric pollution (e.g., Tkacz et

al., 2008) are not relevant in the study area.

89

B. attenuata was the only species whose regeneration was significantly explained in the

study. The higher density of B. attenuata seedlings in the larger remnants is not easily

explained, because the effect of factors known to influence seedling survival and

establishment, including trampling (Matlack, 1993; Lehvävirta and Rita, 2002) and soil

fertility (Aronson and Handel, 2011), and which are related with remnant area (Chapter

3), were not significant. However, other edge effects, including higher seed predation

and alteration of the microclimatic conditions (e.g., higher wind speed and temperature)

(Harper et al., 2005), might explain the lower density of seedlings in the smaller

remnants.

The higher density of B. attenuata saplings in rural and recently urbanized remnants,

where tree density was higher, is relatively intuitive. However, the fact that density of B.

attenuata trees was lower in the older urban remnants, despite a lower mortality in those

remnants, might indicate poor regeneration as well as large past mortality. The fact that

the density of B. attenuata was higher in remnants not burnt for a long period might

indicate that recruitment in some remnants may be hampered by too frequent fires

(Chapter 3), resulting in a lowered seed production (Enright et al., 1998). Another

possible factor that could compromise regeneration in the older and more isolated urban

remnants (Chapter 3) is limited seed dispersal (e.g., Komuro and Koike, 2005;

D’Orangeville et al., 2008). However, this is unlikely because B. attenuata seeds are

large and have low mobility, and only a small percentage is long-distance dispersed (He

et al., 2009).

While density of B. attenuata was lower in the older urban remnants, density of E.

marginata was higher in those remnants. The turnover between Eucalyptus-dominated

and Banksia-dominated woodlands in south-western Australia is influenced by water

90

and soil phosphorus availability, with Eucalyptus woodlands dominating in regions with

higher water and nutrient resources (Lambers et al., 2006). The older urban remnants

are soil nutrient-enriched (Chapter 3), possibly due to runoff from surrounding paved

and fertilized areas, and atmospheric pollution (Kaye et al., 2006; Park et al., 2010).

Although a positive effect of soil fertility was not observed, such an effect has been

observed before on the relative abundance of ectomycorrhizal tree species (including E.

marginata) of the study remnants (Chapter 4). Therefore, higher water and nutrient

availability might explain the higher density of E. marginata in the older urban

remnants. The same reason might also underlie the higher density of B. menziesii in

some of the historically small remnants, as similar edge effects have been described for

this species (Lamont et al., 1994). Finally, it should not be excluded that the density and

frequency of E. marginata, whose timber has a high commercial value, might be shaped

by past logging activities, as has been described in other urban forests (Fahey, 1998).

Final remarks

The results obtained in this study indicated that remnants urbanized for longer,

predominantly those historically small, have a different canopy composition when

compared to remnants in rural areas or that have been recently fragmented. Indeed,

older urban remnants had lower mortality of B. attenuata and B. menziesii, as well as

higher density of E. marginata and B. menziesii live trees. It was suggested that water

and nutrient subsidies due to runoff from surrounding paved, irrigated, and fertilized

areas (Kaye et al., 2006; Koerner and Klopatek, 2010; Park et al., 2010) could explain

those patterns. As other authors have indicated, in regions where resources are limiting,

as is the case for the study area, higher resource availability in urban areas can influence

remnant vegetation (Imhoff et al., 2000; Martin and Stabler, 2002; Kaye et al., 2006). In

cities with mediterranean or semi-arid climates, higher water availability (especially

91

during dry summer months), might increase the plant water status (Martin and Stabler,

2002) and thus might lower drought-induced tree mortality. Coupled with higher

nutrient availability, it may also lead to changes in the canopy composition of the urban

remnants and increased ecosystem productivity (Imhoff et al., 2000). Overall, the

observed changes in the canopy of the older urban remnants are likely to affect the

understorey structure and composition (Ellison et al., 2005; see Chapter 3 and Chapter

4), leading to novel environmental conditions (e.g., lower sunlight) and possibly altered

hydrological and carbon cycles (Breshears et al., 2005).

93

6. CONCLUSIONS

The main underlying theme of this thesis was that an approach explicitly considering

the temporal dynamics of landscape change and the main environmental factors filtering

plant communities in urban remnants might help ecologists to better understand the

effects of contemporary urbanization on those plant communities. The following are the

main contributions and key results of the study.

6.1 Key results

Chapter 2 reviewed the way urbanization has been predominantly assessed in urban

ecological studies, and suggested a new framework to support the study and

conservation of remnant ecosystems in rapidly urbanizing cities. This framework

highlighted how contemporary urbanization differs from previous patterns of urban

expansion, requiring ecologists to acknowledge the phenomenon actively in terms of the

ways that they intervene in, and study, cities. The proposed framework is based on three

major elements: (i) identification of the main environmental factors shaping the

community of study using an ecologically oriented approach; (ii) an explicit temporal

perspective, considering the dynamics of landscape change, and the effects of land-use

legacies and time-lags in the biodiversity response to ongoing environmental change;

(iii) the use of hierarchical approaches as conceptual and analytical tools to understand

relationships between inter-related environmental factors, and their direct and indirect

effects on remnant plant communities.

The framework guided data collection and data analysis in a case study of 30 remnant

Banksia woodlands, located in the relatively young and rapidly expanding city of Perth,

94

in the south-western Australian global biodiversity hotspot (Chapters 3-5). Chapters 3,

4, and 5 examined the effects of a comprehensive set of landscape and local factors on

different aspects of the plant community. Chapter 3 focused on the species richness and

abundance. Chapter 4 focused on the community functional composition with special

attention to five plant functional traits (i.e., growth form, seed dispersal, pollination

model, nutrient acquisition, and regeneration strategy). Finally, Chapter 5 analysed the

population mortality, density, and regeneration of the four most abundant canopy

species. Common patterns emerging from these three chapters include the following:

Plant community response to urbanization-induced fragmentation was time-

lagged, with most of the changes observed along an axis of time since isolation

and only in the smaller remnants (1-5 ha). These results indicate that the impacts

of fragmentation on remnant vegetation of the Perth Metropolitan Area are still

largely yet to be seen and will unfold with time.

Nevertheless, in small remnants (1-5 ha), the plant community changed

rapidly with time since isolation. For instance, small remnants isolated for 45

years or more had nearly half of the native woody and herbaceous plant species

richness than similarly sized remnants recently fragmented. Older small

remnants had also substantially different composition and structure, with a

higher relative abundance of trees, much lower relative abundance of shrubs,

and a changed canopy composition.

Smaller and older urban remnants were more fertile, more productive, with

higher litter accumulation and higher soil organic C, and were more disturbed by

human activities (e.g., trampling). Furthermore, fire frequency and grazing by

native herbivores were largely reduced in the small remnants.

Thus, the changes observed in the plant community of the smaller remnants

are likely to reflect the synergistic effect of multiple fragmentation-related

95

factors. These include the effects on colonization-extinction dynamics, but

predominantly, given the steep slope of some of the observed changes and the

long life span of most species, the effects on the disturbance regimes and local

environmental conditions.

Chapter 3 showed that richness and abundance of woody species were higher in

historically large remnants and lower in the more connected rural city fringes. This

negative effect of landscape connectivity on the woody plant community was tied to the

city development history and past land uses, and highlights the importance of

considering those aspects in the study of urban remnant vegetation (Hahs et al., 2009;

Chapter 2). Chapter 3 also indicated that in the smaller remnants, the richness of native

herbaceous species declined with time since isolation and with soil organic C. An

increased abundance of non-native herbaceous species and litter depth in the older and

smaller remnants may have reduced the abundance of native herbaceous species in those

remnants. Additionally, Chapter 3 suggested that richness of non-native herbaceous

species was higher in the remnants closer to the city centre, possibly due to higher

proximity to the main and older pathways of species introduction (McKinney, 2001;

Hulme, 2009). Finally, Chapter 3 indicated that the abundance of non-native herbaceous

species in the study remnants was strongly influenced by herbivory. Non-native

herbivores (i.e., European rabbit) were associated with higher invasion levels, possibly

because they strongly disturb the topsoil, which predisposes invasion in the south-

western Australian ecosystems (Hopper, 2009). In contrast, native herbivores (i.e.,

western grey kangaroo) were associated with lower invasion levels because they

preferably consume non-native rather than native plant species (Maron and Vilà, 2001;

Levine et al., 2004).

96

The use of a functional trait approach in Chapter 4 provided insight into some of the

changes observed in the smaller remnants, highlighting the importance of such an

approach to scale up patterns in the plant community to the ecosystem level (Suding et

al., 2008). Indeed, Chapter 4 indicated that smaller and older remnants had higher

relative abundance of trees, with a higher proportion of ectomycorrhizal tree species.

Higher nutrient (and possibly water) resource availability in those remnants (Kaye et al.,

2006) may explain the observed results. This shift in the canopy might have led to lower

sunlight available to the understorey, higher litter accumulation, and higher percent of

soil organic C. These changes in the local environment apparently reduced the richness

and abundance of native herbaceous species, as mentioned before, and might explain the

observed steep decline in the relative abundance of shrubs.

Other key results of Chapter 4 for the understorey included that insect-pollinated plant

species declined with time since isolation, which suggests that insect pollinators might

be declining in the urban matrix due to a combination of factors that include limited

immigration and habitat degradation. The relative abundance of bird-pollinated species

was higher in larger remnants, but it did not decrease with time since isolation, which

might reflect the higher mobility of birds and ability of generalist bird-pollinators to

exploit food and habitat resources in the urban matrix. In rural remnants, soil physical

disturbance due to rabbits and a lasting livestock-grazing legacy have probably

contributed to the displacement of the native herbaceous community by non-native

species with wind-dispersed seeds. Nutrient acquisition and regeneration traits in the

understorey seemed to interact with other inter-dependent plant traits not considered in

the study.

97

Finally, Chapter 5 revealed interesting patterns in the canopy of the remnant Banksia

woodlands of the Perth Metropolitan Area, adding insight into some of the canopy-

related patterns observed in the previous chapters. Indeed, Chapter 5 indicated that the

older and historically small urban remnants had lower mortality of the co-dominants B.

attenuata and B. menziesii, as well as higher density of E. marginata trees. Following

evidence from other studies, it was suggested that higher resource availability

originating from neighbouring urban land uses, might lower drought-induced tree

mortality and increase ecosystem productivity (Imhoff et al., 2000; Martin and Stabler,

2002; Kaye et al., 2006; Koerner and Klopatek, 2010). Chapter 5 further indicated that

despite the lower mortality of B. attenuata in the older urban remnants, the density of

live trees and saplings of this species was lower in those remnants. Poor regeneration,

possibly due to high fire frequency, which leads to lower seed production (Enright et

al., 1998), may explain the observed results.

6.2 The effects of urbanization on remnant plant communities –

reflections in a broader context

The effects of urbanization on remnant plant communities are the result of a complex

combination of environmental filters (Williams et al., 2009). The constellation of

factors filtering remnant plant communities is likely not the same across every city, but

depend on three major aspects: (i) the city’s fragmentation and development history; (ii)

the biogeographic and environmental settings; (iii) and the characteristics of the plant

community of interest. Acknowledgement of these aspects is likely to provide a

common ground for the understanding of vegetation patterns obtained across multiple

cities. These aspects are discussed in detail in the following paragraphs.

98

The constellation of factors filtering remnant plant communities in urban areas depends

on a city's fragmentation and development history. In a valuable synthesis about plant

extinction rates in urban areas, Hahs et al. (2009) have distinguished three types of

cities based on their history of landscape transformation:

Cities with extensive landscape transformation prior to 1600 AD; theses cities

have a history of hundreds or even thousands of years of human occupation

through agriculture and pastoral activities (e.g., European cities) (Type I);

Cities with extensive landscape transformation between 1600 AD and 1800

AD. These cities experienced intensive agricultural development prior to, or in

association with, urban expansion (e.g., Chicago, Singapore) (Type II);

Cities with extensive landscape transformation since around 1800 AD and

which are expanding largely into relatively intact natural vegetation (e.g., Cape

Town, Los Angeles).

The city of Perth fits into the Type III category and the results obtained in this study

echo those obtained by Hahs et al. (2009). In Type III cities, habitat fragmentation is

likely to have a major impact on remnant vegetation and explain current and future

patterns in these remnants. Supporting this idea, Hahs et al. (2009) indicated that 67%

of the variation in extinction rates in cities of Type III was explained by the proportion

of extant native vegetation, contrasting with only 2% in cities of Type II. An important

characteristic of Type III cities is that fragmentation impacts are still largely yet to be

seen and hidden in the form of extinction debts. Only with time, those extinction debts

will “be paid” with consequent increases in extinction rates, as is already observed in

cities of Type II (Hahs et al., 2009). Hence, a temporal perspective is crucial to

understand the effects of urbanization-induced fragmentation in Type III cities. In this

respect, one of the important contributions of this thesis is that it has shown that

99

research in the smaller remnants and along an axis of time since remnant isolation

and/or time since urbanization can help ecologists to forecast changes in remnant plant

communities. Furthermore, such an approach provides a way of space for time

substitution (Pickett, 1990) and helps overcome the shortage of historical floristic data

(Cousins, 2009).

In Type I cities the reality is likely to be very different, as “intact vegetation” was

altered long ago, and urbanization is largely expanding into semi-natural vegetation in a

trajectory of recovery following past land uses (e.g., old-fields, regenerated forests)

(Kowarik, 2005). Human occupation through agriculture and pastoral activities have

shaped these plant communities for a long time (Grove and Rackham, 2001; Kowarik,

2005) and have likely led to major extinctions in the past (Hahs et al., 2009).

Nevertheless, in these cities (e.g., surrounding the Mediterranean Basin), plant

communities have already evolved with an agrarian land cover. Indeed, numerous

examples in the literature point to the importance of traditional land management

practices for the conservation of plant communities in that global biodiversity hotspot

(see Grove and Rackham, 2001). Therefore, whereas fragmentation effects might not be

as relevant as in Type III cities, contemporary sprawl of residential and touristic areas

into rural and semi-natural areas might drive future extinctions. Such seems to be

indicated in the results obtained by Hahs et al. (2009), with 39% of the variation in

extinction rates in Type I cities explained by the proportion of extant native vegetation.

The constellation of factors filtering remnant plant communities in cities also depends

on the biogeographic and environmental settings. The predominant lack of specialized

seed-dispersal mechanisms, large dependence on insects and birds for pollination, and

high vulnerability to soil disturbance are characteristics that shape plant responses to

100

urbanization in the ancient and nutrient-impoverished regions of the globe, as in south-

western Australia (Hopper 2009). However, in cities located in younger, post-glacial,

and more fertile regions, different responses to urban environmental filters can be

expected. Also, in cities located in resource-limited regions, urbanization may provide

resource subsidies (water and nutrients) that can influence remnant vegetation,

increasing plant water status (Martin and Stabler, 2002) and ecosystem productivity

(Imhoff et al., 2000). In regions where resources are not limiting, those effects are

unlikely and, in fact, urbanization decreases ecosystem productivity (Imhoff et al.,

2000). As Cuffney et al. (2010) recently suggested, environmental and biogeographic

settings establish the “background conditions upon which urbanization acts and forms

the basis for the variability observed among metropolitan areas in the physical,

chemical, and biological responses to urbanization”.

Finally, the constellation of factors filtering remnant plant communities in cities

depends on the characteristics of the study community. Although this is rather obvious,

it is easily dismissed in studies relying only on the use of aggregated urbanization

measures (Chapter 2).

6.3 Limitations of the study

In this study, an approach based on trying to describe the direct and indirect effects of

different landscape and local factors was adopted. Further, it was suggested that such an

approach should be more widely explored in future urban ecological studies, namely

through structural equation modelling. Nevertheless, caution is necessary in the

interpretation of the observed relationships obtained using these modelling approaches.

Although the results point to causal relationships, they can only do so in a preliminary

101

manner. Clearly, given the inter-related nature of the environmental variables,

disentangling causal links requires more detailed and experimental approaches that were

beyond the scope of this study.

6.4 Future directions

6.4.1 Urbanization as a set of filters

Considering urbanization as a set of filters as a conceptual baseline and to guide in

data collection. Several authors have made a call for a more mechanistic understanding

of the urbanization and fragmentation processes (e.g., Shochat et al., 2006; Didham et

al., 2007; Schlesinger et al., 2008; Williams et al., 2009; Didham et al., 2012). This

may be achieved if urbanization is depicted as a set of multiple inter-related

environmental factors that filter plant communities (Williams et al., 2009). To identify

the main environmental factors relevant in a certain study context, future studies should

consider the fragmentation and development history of the city (Hahs et al., 2009;

Duncan et al., 2011), the type of urban expansion, and the biogeographic and major

environmental settings. A focus on the community or ecological questions of interest is

also necessary, given that the response to the environment is species- and/or trait-

specific (Massol et al., 2011; Schleicher et al., 2011).

Hierarchical approaches in data collection and data analysis. Given the inter-related

and heterarchical nature of factors in urban settings (Crumley, 1994, 2007), hierarchical

approaches provide a way to understand relationships among multiple factors, as well as

their direct and indirect effects on remnant plant communities (Wu and David, 2002;

Qian et al., 2010). Hierarchical approaches available in ecology include the patch

dynamics framework (Wu and Loucks, 1995), structural equation (Grace et al., 2010)

102

and Bayesian hierarchical modelling (McMahon and Diez, 2007; Qian et al., 2010), and

their use should be explored in future studies.

Comparative studies in a controlled multivariate environment. Comparative studies

are an important element of urban ecology research, as they help in the identification of

emergent patterns across multiple cities (McDonnell and Hahs, 2008, 2009; du Toit and

Cilliers, 2011). Future comparative studies should take into account the city’s

development history (Hahs et al., 2009; Duncan et al., 2011), biogeographic settings

(Hopper, 2009), and types of urban expansion (Schneider and Woodcock, 2008), to

guide in the identification of groups of cities that can and cannot be compared or

lumped together.

Urban typologies in comparative studies. In this thesis, two main urban typologies

were used - contemporarily developed cities vs. historically developed cities (Seto et al.,

2011) and cities Type I - III (Hahs et al., 2009). Although these typologies were used to

convey some of the conceptual ideas developed in the thesis, they may lead to

oversimplifications and do not necessarily acknowledge that combinations of types

exist. Future comparative studies should better explore how these typologies intersect

and possibly, develop a more comprehensive classification system considering the cities

age, fragmentation history, and patterns of urban expansion.

Plant functional trait approaches. Environmental factors filter plant communities on

the basis of their plant functional traits (Lavorel and Garnier, 2002; Chapin, 2003;

McGill et al., 2006; Webb et al., 2010). Combined approaches using this framework

together with an approach considering the main environmental factors in play in the

103

study context is likely to enhance current understanding of the effects of urbanization on

remnant plant communities, and should be explored in future studies.

6.4.2 A temporal perspective in the study of rapidly urbanizing landscapes

A temporal perspective in studies of rapidly urbanizing landscapes. Future studies set

in rapidly urbanizing landscapes should adopt a temporal perspective and consider the

dynamics of landscape change in the analysis of the effects of urbanization on remnant

plant communities. Current and historical remnant and landscape configurations, time

since isolation and time since urbanization, and trajectories of landscape change are

variables that can be used for that purpose.

Is it possible to expand the concept of extinction debt? As this study indicated, and as

suggested by Williams et al. (2006), the indirect effects of fragmentation, through

alteration of the natural disturbance regimes and anthropogenic disturbances, are likely

to outweigh the effects of fragmentation on colonization-extinction dynamics in urban

settings, mostly in smaller remnants. This means that estimated extinction debts,

regardless of the method used in their calculation (He and Hubbell, 2011), might

underestimate both the pace of change and number of species committed to extinction in

the urban settings. Future research should look at expanding the concept of extinction

debt to take into account other major drivers of extinction that are inter-related with

remnant area.

104

6.4.3 Supporting urban planning, management, and restoration of urban

remnants in a rapidly urbanizing world

Identification of thresholds in remnant area to set aside in urban planning. Future

research on the relationships among species diversity, remnant area, and time since

isolation is necessary to help in the identification of thresholds in remnant area that

prevent steep extinctions and declines due to synergistic effects of multiple disturbance

drivers.

Prioritization of remnants to set aside for conservation and to restore. Prioritization of

areas for conservation is a major area in conservation planning (Pressey et al., 2007).

Future research should look at using temporal information, namely the time since

remnant fragmentation and land-use history, in prioritization for conservation. Priorities

could be, for instance, those remnants without significant land-use legacies and those

that were recently fragmented.

105

LITERATURE CITED

Aguilar, R., Ashworth, L., Galetto, L., Aizen, M.A., 2006. Plant reproductive

susceptibility to habitat fragmentation: review and synthesis through a meta-analysis.

Ecology Letters 9, 968-980.

Aguilar, R., Quesada, M., Ashworth, L., Herrerias-Diego, Y., Lobo, J., 2008. Genetic

consequences of habitat fragmentation in plant populations: susceptible signals in plant

traits and methodological approaches. Molecular Ecology 17, 5177-5188.

Aiken, L.S., West, S.G., 1991. Multiple regression: testing and interpreting interactions.

SAGE Publications Inc., Thousand Oaks, USA.

Airola, T.M., Buchholz, K., 1984. Species structure and soil characteristics of five urban

forest sites along the New Jersey Palisades. Urban Ecology 8, 149-164.

Alberti, M., 2008. Advances in urban ecology: integrating humans and ecological

processes in urban ecosystems. Springer, New York, USA.

Alberti, M., Marzluff, J.M., Shulenberger, E., Bradley, G., Ryan, C., Zumbrunnen, C.,

2003. Integrating humans into Ecology: opportunities and challenges for studying urban

ecosystems. BioScience 53, 1169-1179.

Andersson, E., Ahrné, K., Pyykönen, M., Elmqvist, T., 2009. Patterns and scale

relations among urbanization measures in Stockholm, Sweden. Landscape Ecology 24,

1331-1339.

Appleyard, S., 1995. The impact of urban development on recharge and groundwater

quality in a coastal aquifer near Perth, Western Australia. Hydrogeology Journal 3, 65-

75.

Arbuckle, J.L., 2010. Amos 19.0.0. Amos Development Corporation, Meadville, USA.

106

Armbruster, W.S., Edwards, M.E., Debevec, E.M., 1994. Floral character displacement

generates assemblage structure of Western Australian Triggerplants (Stylidium).

Ecology 75, 315-329.

Aronson, M.F.J., Handel, S.N., 2011. Deer and invasive plant species suppress forest

herbaceous communities and canopy tree regeneration. Natural Areas Journal 31, 400-

407.

Australian Bureau of Metereology, 2011. Climate statistics for Australian locations:

monthly climate statistics (URL: http://www.bom.gov.au).

Australian Bureau of Statistics, 2011. 3218.0 - Regional Population Growth, Australia,

2009-10 (URL: http://www.abs.gov.au).

Band, L.E., Moore, I.D., 1995. Scale: landscape attributes and geographical information

systems. Hydrological Processes 9, 401-422.

Bastin, L., Thomas, C.D., 1999. The distribution of plant species in urban vegetation

fragments. Landscape Ecology 14, 493-507.

Bates, D., Maechler, M., Bolker, B., 2011. Linear mixed-effects models using S4

classes. University of Wisconsin Press, Madison, USA.

Batty, M., 2008. The size, scale, and shape of cities. Science 319 769-771.

Bell, D.T., 2001. Ecological response syndromes in the flora of south-western Western

Australia: fire resprouters vs reseeders. The Botanical Review 67, 417-440.

Bell, D.T., Plummer, J.A., Taylor, S.K., 1993. Seed germination ecology in south-

western Western Australia. The Botanical Review 59, 24-73.

Bell, D.T., Rokich, D.P., McChesney, C.J., Plummer, J.A., 1995. Effects of

temperature, light and gibberellic acid on the germination of seeds of 43 species native

to Western Australia. Journal of Vegetation Science 6, 797-806.

Bell, T.L., Pate, J.S., 1996. Nitrogen and phosphorus nutrition in mycorrhizal

Epacridaceae of south-west Australia. Annals of Botany 77, 389-397.

107

Bellemare, J., Motzkin, G., Foster, D.R., 2002. Legacies of the agricultural past in the

forested present: an assessment of historical land-use effects on rich mesic forests.

Journal of Biogeography 29, 1401-1420.

Bidwell, S., Attiwill, P.M., Adams, M.A., 2006. Nitrogen availability and weed

invasion in a remnant native woodland in urban Melbourne. Austral Ecology 31, 262-

270.

Blair, R.B., 1996. Land use and avian species diversity along an urban gradient.

Ecological Applications 6, 506-519.

Blakely, T.J., Didham, R.K., 2010. Disentangling the mechanistic drivers of ecosystem-

size effects on species diversity. Journal of Animal Ecology 79, 1204-1214.

Bolger, D.T., Suarez, A.V., Crooks, K.R., Morrison, S.A., Case, T.J., 2000. Arthropods

in urban habitat fragments in southern California: area, age, and edge effects. Ecological

Applications 10, 1230-1248.

Bolund, P., Hunhammar, S., 1999. Ecosystem services in urban areas. Ecological

Economics 29, 293-301.

Boone, C.G., Cadenasso, M.L., Grove, J.M., Schwarz, K., Buckley, G.L., 2009.

Landscape, vegetation characteristics, and group identity in an urban and suburban

watershed: why the 60s matter. Urban Ecosystems 13, 255-271.

Borgström, S.T., Elmqvist, T., Angelstam, P., Alfsen-Norodom, C., 2006. Scale

mismatches in management of urban landscapes. Ecology and Society 11, 16.

Breshears, D.D., Cobb, N.S., Rich, P.M., Price, K.P., Allen, C.D., Balice, R.G.,

Romme, W.H., Kastens, J.H., Floyd, M.L., Belnap, J., Anderson, J.J., Myers, O.B.,

Meyer, C.W., 2005. Regional vegetation die-off in response to global-change-type

drought. Proceedings of the National Academy of Sciences 102, 15144-15148.

Broshot, N., 2007. The influence of urbanization on forest stand dynamics in

Northwestern Oregon. Urban Ecosystems 10, 285-298.

108

Broshot, N., 2011. Mortality and recruitment in an urban forest (Forest Park in Portland,

Oregon) between 1993 and 2003. Urban Ecosystems 14, 553-567.

Brown, E.M., Burbidge, A.H., Dell, J., Edinger, D., Hopper, S.D., Wills, R.T., 1997.

Pollination in Western Australia: a database of animals visiting flowers. Western

Australian Naturalists' Club, Perth, Australia.

Brudvig, L.A., Damschen, E.I., 2011. Land-use history, historical connectivity, and land

management interact to determine longleaf pine woodland understory richness and

composition. Ecography 34, 257-266.

Brundrett, M.C., 2009. Mycorrhizal associations and other means of nutrition of

vascular plants: understanding the global diversity of host plants by resolving

conflicting information and developing reliable means of diagnosis. Plant and Soil 320,

37-77.

Burgman, M.A., Keith, D., Hopper, S.D., Widyatmoko, D., Drilla, C., 2007. Threat

syndromes and conservation of the Australian flora. Biological Conservation 134, 73–

82.

Burton, M.L., Samuelson, L.J., Mackenzie, M.D., 2009. Riparian woody plant traits

across an urban-rural land use gradient and implications for watershed function with

urbanization. Landscape and Urban Planning 90, 42-55.

Cane, J.H., Minckley, R.L., Kervin, L.J., Roulston, T.H., Williams, N.M., 2006.

Complex responses within a desert bee guild (Hymenoptera: Apiformes) to urban

habitat fragmentation. Ecological Applications 16, 632-644.

Canham, C.A., Froend, R.H., Stock, W.D., 2009. Water stress vulnerability of four

Banksia species in contrasting ecohydrological habitats on the Gnangara Mound,

Western Australia. Plant, Cell and Environment 32, 64-72.

109

Carreiro, M.M., Howe, K., Parkhurst, D.F., Pouyat, R.V., 1999. Variation in quality and

decomposability of red oak leaf litter along an urban-rural gradient. Biology and

Fertility of Soils 30, 258-268.

Celesti-Grapow, L., Pyšek, P., Jarošík, V., Blasi, C., 2006. Determinants of native and

alien species richness in the urban flora of Rome. Diversity and Distributions 12, 490-

501.

Chace, J.F., Walsh, J.J., 2006. Urban effects on native avifauna: a review. Landscape

and Urban Planning 74, 46-49.

Chamberlain, D.E., Cannon, A.R., Toms, M.P., 2004. Associations of garden birds with

gradients in garden habitat and local habitat. Ecography 27, 589-600.

Chapin, F.S., 2003. Effects of plant traits on ecosystem and regional processes: a

conceptual framework for predicting the consequences of global change. Annals of

Botany 91, 455-463.

Comita, L.S., Thompson, J., Uriarte, M., Jonckheere, I., Canham, C.D., Zimmerman,

J.K., 2010. Interactive effects of land use history and natural disturbance on seedling

dynamics in a subtropical forest. Ecological Applications 20, 1270-1284.

Commander, P., Hauck, E., 2005. How our groundwater has changed. Indian Ocean

Climate Initiative (Western Australian Government, CSIRO and Bureau of

Meteoreology), Perth, Australia.

Cousins, S.A.O., 2009. Extinction debt in fragmented grasslands: paid or not?

International Association for Vegetation Science 20, 3-7.

Cousins, S.A.O., Ohlson, H., Eriksson, O., 2007. Effects of historical and present

fragmentation on plant species diversity in semi-natural grasslands in Swedish rural

landscapes. Landscape Ecology 22, 723-730.

110

Crosti, R., Dixon, K.W., Ladd, P.G., Yates, C.J., 2007. Generic changes in the structure

and species dominance in vegetation over 60 years in an urban bushland remnant

Pacific Conservation Biology 13, 158-170.

Crumley, C.L., 1994. Historical Ecology: a multidimensional ecological orientation. In:

Crumley, C.L. (Ed.), Historical ecology: cultural knowledge and changing landscapes.

School of American Research Press, Santa Fe, pp. 1-16.

Crumley, C.L., 2007. Historical ecology: integrated thinking at multiple temporal and

spatial scales. In: Hornborg, A., Crumley, C.L. (Eds.), The world system and the Earth

system: global socio-environmental change and sustainability since the Neolithic. Left

Coast Press, Walnut Creek, USA, pp. 15-28.

Cuffney, T.F., Brightbill, R.A., May, J.T., Waite, I.R., 2010. Responses of benthic

macroinvertebrates to environmental changes associated with urbanization in nine

metropolitan areas. Ecological Applications 20, 1384-1401.

Czech, B., Krausman, P., Devers, P., 2000. Economic associations among causes of

species endangerment in the United States. BioScience 50, 593-601.

D’Orangeville, L., Bouchard, A., Cogliastro, A., 2008. Post-agricultural forests:

Landscape patterns add to stand-scale factors in causing insufficient hardwood

regeneration. Forest Ecology and Management 255, 1637-1646.

Dambrine, E., Dupouey, J.-L., Laüt, L., Humbert, L., Thinon, M., Beaufils, T., Richard,

H., 2007. Present forest biodiversity patterns in France related to former Roman

agriculture. Ecology 88, 1430–1439.

Damschen, E.I., Brudvig, L.A., Haddad, N.M., Levey, D.J., Orrock, J.L., Tewksbury,

J.J., 2008. The movement ecology and dynamics of plant communities in fragmented

landscapes. Proceedings of the National Academy of Sciences 105, 19078-19083.

Dell, B., Bartle, J., Tacey, W., 1983. Root occupation and root channels of jarrah forest

subsoils. Australian Journal of Botany 31, 615-627.

111

Díaz, S., Cabido, M., Casanoves, F., 1999. Functional implications of trait–environment

linkages in plant communities. In: E. Weiher, Keddy, P. (Eds.), Ecological Assembly

Rules – Perspectives, Advances, Retreats. Cambridge University Press, Cambridge, UK,

pp. 338-362.

Didham, R.K., Kapos, V., Ewers, R.M., 2012. Rethinking the conceptual foundations of

habitat fragmentation research. Oikos 121, 161-170.

Didham, R.K., Tylianakis, J.M., Gemmell, N.J., Rand, T.A., Ewers, R.M., 2007.

Interactive effects of habitat modification and species invasion on native species

decline. Trends in Ecology and Evolution 22, 489-496.

Dodd, J., Bell, D.T., 1993. Water relations of understorey shrubs in a Banksia

woodland, Swan Coastal Plain, Western Australia. Australian Journal of Botany 18,

295-305.

Dodd, J., Griffin, E.A., 1989. Floristics of the Banksia woodlands. Journal of the Royal

Society of Western Australia 71, 89-90.

Dray, S., Dufour, A.B., 2007. The ade4 package: implementing the duality diagram for

ecologists. Journal of Statistical Software 22, 1-20.

Dray, S., Legendre, P., 2008. Testing the species traits–environment relationships: the

fourth-corner problem revisited. Ecology 89, 3400-3412.

Dray, S., Legendre, P., Blanchet, F.G., 2007. Packfor: forward selection with

permutation (http://r-forge.r-project.org/R/?group_id=195).

du Toit, M., Cilliers, S., 2011. Aspects influencing the selection of representative

urbanization measures to quantify urban–rural gradients. Landscape Ecology 26, 169-

181.

Duguay, S., Eigenbrod, F., Fahrig, L., 2007. Effects of surrounding urbanization on

non-native flora in small forest patches. Landscape Ecology 22, 589-599.

112

Duncan, R.P., Clemants, S.E., Corlett, R.T., Hahs, A.K., McCarthy, M.A., McDonnell,

M.J., Schwartz, M.W., Thompson, K., Vesk, P.A., Williams, N.S.G., 2011. Plant traits

and extinction in urban areas: a meta-analysis of 11 cities. Global Ecology and

Biogeography 20, 509-519.

Duncan, R.P., Young, J.R., 2000. Determinants of plant extinction and rarity 145 years

after european settlement of Auckland, New Zealand. Ecology 81, 3048-3061.

Dunn, R.R., Gavin, M.C., Sanchez, M.C., Solomon, J.N., 2006. The pigeon paradox:

dependence of global conservation on urban nature. Conservation Biology 20, 1814-

1816.

Ellison, A.M., Bank, M.S., Clinton, B.D., Colburn, E.A., Elliott, K., Ford, C.R., Foster,

D.R., Kloeppel, B.D., Knoepp, J.D., Lovett, G.M., Mohan, J., Orwig, D.A.,

Rodenhouse, N.L., Sobczak, W.V., Stinson, K.A., Stone, J.K., Swan, C.M., Thompson,

J., Von Holle, B., Webster, J.R., 2005. Loss of foundation species: consequences for the

structure and dynamics of forested ecosystems. Frontiers in Ecology and the

Environment 3, 479-486.

Enright, N.J., Marsula, R., Lamont, B.B., Wissel, C., 1998. The ecological significance

of canopy seed storage in fire-prone environments: a model for resprouting shrubs.

Journal of Ecology 86, 960-973.

ESRI, 2008. ArcGIS 9.3. Environmental Systems Research Institute, Redlands, USA.

Esseen, P.-A., 1994. Tree mortality patterns after experimental fragmentation of an old-

growth conifer forest. Biological Conservation 68, 19-28.

Essl, F., Dullinger, S., Rabitsch, W., Hulme, P.E., Hülber, K., Jarošík, V., Kleinbauer,

I., Krausmann, F., Kühn, I., Nentwig, W., Vilà, M., Genovesi, P., Gherardi, F., Desprez-

Loustau, M.-L., Roques, A., Pyšek, P., 2011. Socioeconomic legacy yields an invasion

debt. Proceedings of the National Academy of Sciences 108, 203-207.

113

Ewers, R.M., Didham, R.K., 2006. Confounding factors in the detection of species

responses to habitat fragmentation. Biological Reviews 81, 117-142.

Facelli, J., Pickett, S., 1991. Plant litter: its dynamics and effects on plant community

structure. The Botanical Review 57, 1-32.

Fahey, T.J., 1998. Recent changes in an upland forest in south-central New York.

Journal of the Torrey Botanical Society 125, 51-59.

Fisher, J.L., Loneragan, W.A., Dixon, K., Delaney, J., Veneklaas, E.J., 2009a. Altered

vegetation structure and composition linked to fire frequency and plant invasion in a

biodiverse woodland. Biological Conservation 142, 2270-2281.

Fisher, J.L., Loneragan, W.A., Dixon, K., Veneklaas, E.J., 2009b. Soil seed bank

compositional change constrains biodiversity in an invaded species-rich woodland.

Biological Conservation 142, 256-269.

Fitzpatrick, M.C., Gove, A.D., Sanders, N.J., Dunn, R.R., 2008. Climate change, plant

migration, and range collapse in a global biodiversity hotspot: the Banksia (Proteaceae)

of Western Australia. Global Change Biology 14, 1337-1352.

Flinn, K.M., Vellend, M., 2005. Recovery of forest plant communities in post-

agricultural landscapes. Frontiers in Ecology and the Environment 3, 243-250.

Foster, D., Swanson, F., Aber, J., Burke, I., Brokaw, N., Tilman, D., Knapp, A., 2003.

The importance of land-use legacies to ecology and conservation. BioScience 53, 77-88.

Froend, R.H., Drake, P.L., 2006. Defining phreatophyte response to reduced water

availability: preliminary investigations on the use of xylem cavitation vulnerability in

Banksia woodland species. Australian Journal of Botany 54, 173-179.

Fuller, R.A., Irvine, K.N., Devine-Wright, P., Warren, P.H., Gaston, K.J., 2007.

Psychological benefits of greenspace increase with biodiversity. Biology Letters 3, 390-

394.

114

Garden, J., McAlpine, C., Possingham, H., 2010. Multi-scaled habitat considerations for

conserving urban biodiversity: native reptiles and small mammals in Brisbane,

Australia. Landscape Ecology 25, 1013-1028.

Garnier, E., Cortez, J., Billès, G., Navas, M.-L., Roumet, C., Debussche, M., Laurent,

G., Blanchard, A., Aubry, D., Bellmann, A., Neill, C., Toussaint, J.-P., 2004. Plant

functional markers capture ecosystem properties during secondary sucession. Ecology

85, 2630-2637.

Gavier-Pizarro, G.I., Radeloff, V.C., Stewart, S.I., Huebner, C.D., Keuler, N.S., 2010.

Housing is positively associated with invasive exotic plant species richness in New

England, USA. Ecological Applications 20, 1913-1925.

Germaine, S.S., Wakeling, B.F., 2001. Lizard species distributions and habitat

occupation along an urban gradient in Tucson, Arizona, USA. Biological Conservation,

229-237.

Gillson, L., 2009. Landscapes in time and space. Landscape Ecology 24, 149-155.

Goddard, M.A., Dougill, A.J., Benton, T.G., 2010. Scaling up from gardens:

biodiversity conservation in urban environments. Trends in Ecology and Evolution 25,

90-98.

Godefroid, S., Koedam, N., 2003. How important are large vs. small forest remnants for

the conservation of the woodland flora in an urban context? Global Ecology and

Biogeography 12, 287-298.

Godefroid, S., Koedam, N., 2007. Urban plant species patterns are highly driven by

density and function of built-up areas. Landscape Ecology 22, 1227-1239.

Grace, J.B., 2006. Structural equation modeling and natural systems. Cambridge

University Press, Cambridge, UK.

Grace, J.B., Anderson, T.M., Olff, H., Scheiner, S.M., 2010. On the specification of

structural equation models for ecological systems. Ecological Monographs 80, 67-87.

115

Grigg, A.M., Pate, J.S., Unkovich, M.J., 2000. Responses of native woody taxa in

Banksia woodland to incursion of groundwater and nutrients from bordering agricultural

land. Australian Journal of Botany 48, 777-792.

Grimm, N.B., Faeth, S.H., Golubiewski, N.E., Redman, C.L., Wu, J., Bai, X., Briggs,

J.M., 2008. Global change and the ecology of cities. Science 319, 756-760.

Grimm, N.B., Grove, J.M., Pickett, S.T.A., Redman, C.L., 2000. Integrated approaches

to long-term studies of urban ecological systems. BioScience 50, 571-584.

Groom, B.P.K., Froend, R.H., Mattiske, E.M., 2000. Impact of groundwater abstraction

on a Banksia woodland, Swan Coastal Plain, Western Australia. Ecological

Management and Restoration 1, 117-124.

Grove, A.T., Rackham, O., 2001. The nature of Mediterranean Europe - an ecological

history. Yale University Press, New Haven.

Guirado, M., Pino, J., Rodà, F., 2006. Understorey plant species richness and

composition in metropolitan forest archipelagos: effects of forest size, adjacent land use

and distance to the edge. Global Ecology and Biogeography 15, 50-62.

Guntenspergen, G., Levenson, J., 1997. Understory plant species composition in

remnant stands along an urban-to-rural land-use gradient. Urban Ecosystems 1, 155-

169.

Guyette, R.P., Muzika, R.M., Dey, D.C., 2002. Dynamics of an anthropogenic fire

regime. Ecosystems 5, 472-486.

Hahs, A., McDonnell, M., 2007. Composition of the plant community in remnant

patches of grassy woodland along an urban–rural gradient in Melbourne, Australia.

Urban Ecosystems 10, 355-377.

Hahs, A.K., McDonnell, M.J., 2006. Selecting independent measures to quantify

Melbourne’s urban-rural gradient. Landscape and Urban Planning 78, 435-448.

116

Hahs, A.K., McDonnell, M.J., McCarthy, M.A., Vesk, P.A., Corlett, R.T., Norton, B.A.,

Clemants, S.E., Duncan, R.P., Thompson, K., Schwartz, M.W., Williams, N.S.G., 2009.

A global synthesis of plant extinction rates in urban areas. Ecology Letters 12, 1165-

1173.

Hamberg, L., Lehvävirta, S., Malmivaara-Lämsä, M., Rita, H., Kotze, D.J., 2008. The

effects of habitat edges and trampling on understorey vegetation in urban forests in

Helsinki, Finland. Applied Vegetation Science 11, 83-98.

Hanski, I., 1998. Metapopulation dynamics. Nature 396, 41-49.

Harper, K.A., Macdonald, S.E., Burton, P.J., Chen, J., Brosofske, K.D., Saunders, S.C.,

Euskirchen, E.S., Roberts, D.A.R., Jaiteh, M.S., Esseen, P.-A., 2005. Edge influence on

forest structure and composition in fragmented landscapes. Conservation Biology 19,

768-782.

Hastings, A., 2010. Timescales, dynamics, and ecological understanding. Ecology 91,

3471-3480.

He, F., Hubbell, S.P., 2011. Species-area relationships always overestimate extinction

rates from habitat loss. Nature 473, 368-371.

He, T., Lamont, B.B., Krauss, S.L., Enright, N.J., Miller, B.P., 2009. Long-distance

dispersal of seeds in the fire-tolerant shrub Banksia attenuata. Ecography 32, 571-580.

Hedblom, M., Söderström, B., 2010. Landscape effects on birds in urban woodlands: an

analysis of 34 Swedish cities. Journal of Biogeography 37, 1302-1316.

Helm, A., Hanski, I., Partel, M., 2006. Slow response of plant species richness to habitat

loss and fragmentation. Ecology Letters 9, 72-77.

Henle, K., Davies, K.F., Kleyer, M., Margules, C., Settele, J., 2004. Predictors of

species sensitivity to fragmentation. Biodiversity and Conservation 13, 207-251.

Hill, T.C.J., Tippett, J.T., Shearer, B.L., 1994. Invasion of Bassendean dune Banksia

woodland by Phytophthora cinnamomi. Australian Journal of Botany 42, 725-738.

117

Hobbs, R., Atkins, L., 1990. Fire-related dynamics of a Banksia woodland in south-

western Australia. Australian Journal of Botany 38, 97-110.

Hobbs, R.J., Huenneke, L.F., 1992. Disturbance, diversity, and invasion: implications

for conservation. Conservation Biology 6, 324-337.

Hobbs, R.J., Yates, C.J., 2003. Impacts of ecosystem fragmentation on plant

populations: generalising the idiosyncratic. Australian Journal of Botany 51, 471-488.

Holm, E., 1988. On pollination and pollinators in Western Australia. Denmark.

Holm, S., 1979. A Simple sequentially rejective multiple test procedure. Scandinavian

Journal of Statistics 6, 65-70.

Hopper, S., 2009. OCBIL theory: towards an integrated understanding of the evolution,

ecology and conservation of biodiversity on old, climatically buffered, infertile

landscapes. Plant and Soil 322, 49-86.

Hopper, S.D., Gioia, P., 2004. The Southwest Australian Floristic Region: conservation

of a global hotspot of biodiversity. Annual Review of Ecology and Systematics 35, 623-

650.

Hostetler, M., Holling, C.S., 2000. Detecting the scales at which birds respond to

structure in urban landscapes. Urban Ecosystems 4, 25-54.

Hulme, P.E., 2009. Trade, transport and trouble: managing invasive species pathways in

an era of globalization. Journal of Applied Ecology 46, 10-18.

Hutton, B.J., Dixon, K.W., Sivasithamparam, K., Pate, J.S., 1996. Inoculum potential of

ericoid endophytes of Western Australian heaths (Epacridaceae). New Phytologist 134,

665-672.

Hutton, B.J., Dixon, K.W., Sivasithamparam, K., Pate, J.S., 1997. Effect of habitat

disturbance on inoculum potential of ericoid endophytes of Western Australian heaths

(Epacridaceae). New Phytologist 135, 739-744.

118

Iacono, M., Levinson, D., El-Geneidy, A., 2008. Models of transportation and land use

change: a Guide to the territory. Journal of Planning Literature 22, 323-340.

Imhoff, M.L., Tucker, C.J., Lawrence, W.T., Stutzer, D.C., 2000. The use of

multisource satellite and geospatial data to study the effect of urbanization on primary

productivity in the United States. IEEE Transactions on Geoscience and Remote

Sensing 38, 2549-2556.

Jackson, S.T., Sax, D.F., 2010. Balancing biodiversity in a changing environment:

extinction debt, immigration credit and species turnover. Trends in Ecology and

Evolution 25, 153-160.

Jacobsen, A.L., Pratt, R.B., Ewers, F.W., Davis, S.D., 2007. Cavitation resistance

among 26 chaparral species of southern California. Ecological Monographs 77, 99-115.

Jim, C.Y., Chen, W.Y., 2009. Ecosystem services and valuation of urban forests in

China. Cities 26, 187-194.

Kaye, J.P., Groffman, P.M., Grimm, N.B., Baker, L.A., Pouyat, R.V., 2006. A distinct

urban biogeochemistry? Trends in Ecology and Evolution 21, 192-199.

Kaye, J.P., McCulley, R.L., Burke, I.C., 2005. Carbon fluxes, nitrogen cycling, and soil

microbial communities in adjacent urban, native and agricultural ecosystems. Global

Change Biology 11, 575-587.

Keighery, G.J., 1980. Bird pollination in the south-western Australia: a checklist. Plant

Systematics and Evolution 135, 171-176.

Keighery, G.J., 1982. Reproductive strategies of Western Australian Apiaceae. Plant

Systematics and Evolution 140, 243-250.

Keighery, G.J., 1996. Phytogeography, biology and conservation of Western Australian

Epacridaceae. Annals of Botany 77, 347-355.

119

Kendrick, G.W., Wyrwoll, K.-H., Szabo, B.J., 1991. Pliocene-Pleistocene coastal events

and history along the western margin of Australia. Quaternary Science Reviews 10,

419-439.

Kennewell, C., Shaw, B.J., 2008. Perth, Western Australia. Cities 25, 243-255.

Knapp, S., Kuhn, I., Stolle, J., Klotz, S., 2010. Changes in the functional composition of

a Central European urban flora over three centuries. Perspectives in Plant Ecology,

Evolution and Systematics 12, 235-244.

Knapp, S., Kühn, I., Wittig, R., Ozinga, W.A., Poschlod, P., Klotz, S., 2008.

Urbanization causes shifts in species' trait state frequencies. Preslia 80, 375-388.

Koerner, B., Klopatek, J., 2010. Carbon fluxes and nitrogen availability along an urban–

rural gradient in a desert landscape. Urban Ecosystems 13, 1-21.

Komuro, T., Koike, F., 2005. Colonization by woody plants in fragmented habitats of a

suburban landscape. Ecological Applications 15, 662-673.

Kostel-Hughes, F., Young, T.P., Carreiro, M.M., 1998. Forest leaf litter quantity and

seedling occurrence along an urban-rural gradient. Urban Ecosystems 2, 263-278.

Kowarik, I., 2005. Wild urban woodlands: towards a conceptual framework. In:

Kowarik, I., Stefan, K. (Eds.), Wild urban woodlands: new perspectives for urban

forestry. Springer-Verlag, Berlin, Germany, pp. 1-32.

Krauss, J., Bommarco, R., Guardiola, M., Heikkinen, R.K., Helm, A., Kuussaari, M.,

Lindborg, R., Ockinger, E., Partel, M., Pino, J., Poyry, J., Raatikainen, K.M., Sang, A.,

Stefanescu, C., Teder, T., Zobel, M., Steffan-Dewenter, I., 2010. Habitat fragmentation

causes immediate and time delayed biodiversity loss at different trophic levels. Ecology

Letters 13, 597-605.

Kuussaari, M., Bommarco, R., Heikkinen, R.K., Helm, A., Krauss, J., Lindborg, R.,

Öckinger, E., Pärtel, M., Pino, J., Rodà, F., Stefanescu, C., Teder, T., Zobel, M.,

120

Steffan-Dewenter, I., 2009. Extinction debt: a challenge for biodiversity conservation.

Trends in Ecology and Evolution 24, 564-571.

Laliberté, E., Legendre, P., 2010. A distance-based framework for measuring functional

diversity from multiple traits. Ecology 91, 299-305.

Lambers, H., Brundrett, M., Raven, J., Hopper, S., 2010. Plant mineral nutrition in

ancient landscapes: high plant species diversity on infertile soils is linked to functional

diversity for nutritional strategies. Plant and Soil 334, 11-31.

Lambers, H., Raven, J.A., Shaver, G.R., Smith, S.E., 2008. Plant nutrient-acquisition

strategies change with soil age. Trends in Ecology and Evolution 23, 95-103.

Lambers, H., Shane, M.W., Cramer, M.D., Pearse, S.J., Veneklaas, E.J., 2006. Root

structure and functioning for efficient acquisition of phosphorus: matching

morphological and physiological traits. Annals of Botany 98, 693-713.

Lamont, B., 1982. Mechanisms for enhancing nutrient uptake in plants, with particular

reference to mediterranean South Africa and Western Australia. The Botanical Review

48, 597-589.

Lamont, B.B., Groom, P.K., 1998. Seed and seedling biology of the woody-fruited

Proteaceae. Australian Journal of Botany 46, 387-406.

Lamont, B.B., Written, V.A., Witkowski, E.T.F., Rees, R.G., Enright, N.J., 1994.

Regional and local (road verge) effects on size and fecundity in Banksia menziesii.

Australian Journal of Ecology 19, 197-205.

Laurance, W.F., Ferreira, L.V., Merona, J.M.R.-D., Laurance, S.G., Hutchings, R.W.,

Lovejoy, T.E., 1998a. Effects of forest fragmentation on recruitment patterns in

Amazonian tree communities. Conservation Biology 12, 460-464.

Laurance, W.F., Ferreira, L.V., Rankin-de Merona, J.M., Laurance, S.G., 1998b. Rain

forest fragmentation and the dynamics of Amazonian tree communities. Ecology 79,

2032-2040.

121

Lavorel, S., Garnier, E., 2002. Predicting changes in community composition and

ecosystem functioning from plant traits: revisiting the Holy Grail. Functional Ecology

16, 545-556.

Lavorel, S., Grigulis, K., McIntyre, S., Williams, N.S.G., Garden, D., Dorrough, J.,

Berman, S., Quétier, F., Thébault, A., Bonis, A., 2008. Assessing functional diversity in

the field – methodology matters! Functional Ecology 22, 134-147.

Legendre, P., Galzin, R., Harmelin-Vivien, M.L., 1997. Relating behavior to habitat:

solutions to the fourth-corner problem. Ecology 78, 547-562.

Legendre, P., Legendre, L., 1998. Numerical ecology. Elsevier Science, Amsterdam,

Netherlands.

Lehvävirta, S., Rita, H., 2002. Natural regeneration of trees in urban woodlands. Journal

of Vegetation Science 13, 57-66.

Lerman, S.B., Warren, P.S., 2011. The conservation value of residential yards: linking

birds and people. Ecological Applications 21, 1327-1339.

Levine, J.M., Adler, P.B., Yelenik, S.G., 2004. A meta-analysis of biotic resistance to

exotic plant invasions. Ecology Letters 7, 975-989.

Lewis, D.B., Kaye, J.P., Gries, C., Kinzig, A.P., Redman, C.L., 2006. Agrarian legacy

in soil nutrient pools of urbanizing arid lands. Global Change Biology 12, 703-709.

Lim, H.C., Sodhi, N.S., 2004. Responses of avian guilds to urbanisation in a tropical

city. Landscape and Urban Planning 66, 199-215.

Liu, J., Dietz, T., Carpenter, S.R., Alberti, M., Folke, C., Moran, E., Pell, A.N.,

Deadman, P., Kratz, T., Lubchenco, J., Ostrom, E., Ouyang, Z., Provencher, W.,

Redman, C.L., Schneider, S.H., Taylor, W.W., 2007. Complexity of coupled human and

natural systems. Science 317, 1513-1516.

122

Lososová, Z., Chytrý, M., Kühn, I., Hájek, O., Horáková, V., Pysek, P., Tichý, L., 2006.

Patterns of plant traits in annual vegetation of man-made habitats in central Europe.

Perspectives in Plant Ecology, Evolution and Systematics 8, 69-81.

Loss, S.R., Ruiz, M.O., Brawn, J.D., 2009. Relationships between avian diversity,

neighborhood age, income, and environmental characteristics of an urban landscape.


Lu, D., Weng, Q., 2006. Use of impervious surface in urban land-use classification.

Remote Sensing of Environment 102, 146-160.

Luck, G.W., 2007. A review of the relationships between human population density and

biodiversity. Biological Reviews 82, 607-645.

Luck, G.W., Smallbone, L.T., O’Brien, R., 2009. Socio-economics and vegetation

change in urban ecosystems: patterns in space and time. Ecosystems 12, 604-620.

Luck, M., Wu, J., 2002. A gradient analysis of urban landscape pattern: a case study

from the Phoenix metropolitan region, Arizona, USA. Landscape Ecology 17, 327-339.

MacArthur, R.H., Wilson, E.O., 1967. The theory of island biogeography. Princeton

University Press, Princeton, NJ, USA.

Machlis, G.E., Force, J.E., Burch, W.R.J., 1997. The human ecosystem Part I: the

human ecosystem as an organizing concept in ecosystem management. Society and

Natural Resources 10, 347-367.

Magle, S.B., Crooks, K.R., 2009. Investigating the distribution of prairie dogs in an

urban landscape. Animal Conservation 12, 192-203.

Magura, T., Horváth, R., Tóthmérész, B., 2010. Effects of urbanization on ground-

dwelling spiders in forest patches, in Hungary. Landscape Ecology 25, 621-629.

Makse, H.A., Havlin, S., Stanley, H.E., 1995. Modeling urban growth patterns. Nature

377, 608-612.

123

Maron, J.L., Vilà, M., 2001. When do herbivores affect plant invasion? Evidence for the

natural enemies and biotic resistance hypotheses. Oikos 95, 361-373.

Martin, C.A., Stabler, L.B., 2002. Plant gas exchange and water status in urban desert

landscapes. Journal of Arid Environments 51, 235-254.

Martínez-Vilalta, J., Prat, E., Oliveras, I., Piñol, J., 2002. Xylem hydraulic properties of

roots and stems of nine Mediterranean woody species. Oecologia 133, 19-29.

Massol, F., Gravel, D., Mouquet, N., Cadotte, M.W., Fukami, T., Leibold, M.A., 2011.

Linking community and ecosystem dynamics through spatial ecology. Ecology Letters

14, 313-323.

Matlack, G., 1993. Sociological edge effects: Spatial distribution of human impact in

suburban forest fragments. Environmental Management 17, 829-835.

May, F., Ash, J., 1990. An assessment of the allelopathic potential of Eucalyptus.

Australian Journal of Botany 38, 245-254.

McArthur, W.M., Johnston, D.A.W., Snell, L.J., 2004. Reference soils of south-western

Australia. Western Australian Department of Agriculture, Perth, Australia.

Mccredie, T., Dixon, K., Sivasithamparam, K., 1985. Variability in the resistance of

Banksia L.f. species to Phytophthora cinnamomi Rands. Australian Journal of Botany

33, 629-637.

McDonald, R.I., Kareiva, P., Forman, R.T.T., 2008. The implications of current and

future urbanization for global protected areas and biodiversity conservation. Biological

Conservation 141, 1695-1703.

McDonnell, M.J., Hahs, A.K., 2008. The use of gradient analysis studies in advancing

our understanding of the ecology of urbanizing landscapes: current status and future

directions. Landscape Ecology 23, 1143-1155.

McDonnell, M.J., Hahs, A.K., 2009. Ecology of cities and towns. A comparative

approach. Cambridge University Press, Cambridge, UK.

124

McDonnell, M.J., Pickett, S.T.A., 1990. Ecosystem structure and function along urban-

rural gradients: an unexploited opportunity for Ecology. Ecology 71, 1232-1237.

McGill, B.J., Enquist, B.J., Weiher, E., Westoby, M., 2006. Rebuilding community

ecology from functional traits. Trends in Ecology and Evolution 21, 178-185.

McIntyre, N.E., Knowles-Yanez, K., Hope, D., 2000. Urban ecology as an

interdisciplinary field: differences in the use of “urban” between the social and natural

sciences. Urban Ecosystems 4, 5-24.

McKey, D., Rostain, S., Iriarte, J., Glaser, B., Birk, J.J., Holst, I., Renard, D., 2010. Pre-

Columbian agricultural landscapes, ecosystem engineers, and self-organized patchiness

in Amazonia. Proceedings of the National Academy of Sciences 107, 7823-7828.

McKinney, M.L., 2001. Effects of human population, area, and time on non-native plant

and fish diversity in the United States. Biological Conservation 100, 243-252.

McKinney, M.L., 2002. Influence of settlement time, human population, park shape and

age, visitation and roads on the number of alien plant species in protected areas in the

USA. Diversity and Distributions 8, 311-318.

McKinney, M.L., 2006. Urbanization as a major cause of biotic homogenization.


McKinney, M.L., 2008. Effects of urbanization on species richness: a review of plants

and animals. Urban Ecosystems 11, 161-176.

McMahon, S.M., Diez, J.M., 2007. Scales of association: hierarchical linear models and

the measurement of ecological systems. Ecology Letters 10, 437-452.

Meney, K.A., Nielssen, G.M., Dixon, K.W., 1994. Seed bank patterns in Restionaceae

and Epacridaceae after wildfire in kwongan in southwestern Australia. Journal of

Vegetation Science 5, 5-12.

125

Meney, K.A., Pate, J.S., Dixon, K.W., 1990. Comparative morphology, anatomy,

phenology and reproductive biology of Alexgeorgea spp. (Restionaceae) from south-

western Australia. Australian Journal of Botany 38, 523-541.

Mesquita, R.C.G., Delamônica, P., Laurance, W.F., 1999. Effect of surrounding

vegetation on edge-related tree mortality in Amazonian forest fragments. Biological

Conservation 91, 129-134.

Miller, J.R., 2005. Biodiversity conservation and the extinction of experience. Trends in

Ecology and Evolution 20, 430-434.

Miller, J.R., Hobbs, R.J., 2002. Conservation where people live and work. Conservation

Biology 16, 330-337.

Moilanen, A., Nieminen, M., 2002. Simple connectivity measures in spatial ecology.

Ecology 83, 1131-1145.

Montgomery, M.R., 2008. The urban transformation of the developing world. Science

319, 761-764.

Morrison, W.E., Hay, M.E., 2011. Herbivore preference for native vs. exotic plants:

generalist herbivores from multiple continents prefer exotic plants that are

evolutionarily naïve. PLoS ONE 6, e17227.

Myers, N., Mittermeier, R.A., Mittermeier, C.G., Fonseca, G.A.B., Kent, J., 2000.

Biodiversity hotspots for conservation priorities. Nature 403, 803-808.

Nelson, K.C., Palmer, M.A., Pizzuto, J.E., Moglen, G.E., Angermeier, P.L.,

Hilderbrand, R.H., Dettinger, M., Hayhoe, K., 2009. Forecasting the combined effects

of urbanization and climate change on stream ecosystems: from impacts to management

options. Journal of Applied Ecology 46, 154-163.

Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., O'Hara, R.B., Simpson, G.L.,

Solymos, P., Stevens, M.H.H., Wagner, H.H., 2011. Vegan: community ecology

package. R package version 1.17-11. http://CRAN.R-project.org/package=vegan.

126

Ovaskainen, O., Hanski, I., 2002. Transient dynamics in metapopulation response to

perturbation. Theoretical Population Biology 61, 285-295.

Pakeman, R.J., 2004. Consistency of plant species and trait responses to grazing along a

productivity gradient: a multi-site analysis. Journal of Ecology 92, 893-905.

Paltto, H., Nordén, B., Götmark, F., Franc, N., 2006. At which spatial and temporal

scales does landscape context affect local density of Red Data Book and Indicator

species? Biological Conservation 133, 442-454.

Park, S.-J., Cheng, Z., Yang, H., Morris, E., Sutherland, M., McSpadden Gardener, B.,

Grewal, P., 2010. Differences in soil chemical properties with distance to roads and age

of development in urban areas. Urban Ecosystems 13, 483-497.

Parker, J.D., Burkepile, D.E., Hay, M.E., 2006. Opposing effects of native and exotic

herbivores on plant invasions. Science 311, 1459-1461.

Parker, J.D., Richie, L.J., Lind, E.M., Maloney, K.O., 2010. Land use history alters the

relationship between native and exotic plants: the rich don’t always get richer.

Biological Invasions 12, 1557-1571.

Pate, J.S., Beard, J.S., 1984. Kwongan, plant life of the Sandplain: biology of a south-

west Australian shrubland ecosystem. University of Western Australia Press, Nedlands,

Australia.

Pate, J.S., Bell, T.L., 1999. Application of the ecosystem mimic concept to the species-

rich Banksia woodlands of Western Australia. Agroforestry Systems 45, 303–341.

Pate, J.S., Froend, R.H., Bowen, B.J., Hansen, A., Kuo, J., 1990. Seedling growth and

storage characteristics of seeder and resprouter species of mediterranean-type

ecosystems of S.W. Australia. Annals of Botany 65, 585-601.

Pate, J.S., Meney, K.A., Dixon, K.W., 1991. Contrasting growth and morphological

characteristics of fire-sensitive (obligate seeder) and fire-resistant (resprouter) species of

127

Restionaceae (S. Hemisphere Restiads) from south-western Western Australia.


Pausas, J.G., Bradstock, R.A., Keith, D.A., Jon, E.K., The, G.F.N., 2004. Plant

functional traits in relation to fire in crown-fire ecosystems. Ecology 85, 1085-1100.

Pauw, A., 2007. Collapse of a pollination web in small conservation areas. Ecology 88,

1759-1769.

Pekin, B.K., Wittkuhn, R.S., Boer, M.M., Macfarlane, C., Grierson, P.F., 2011. Plant

functional traits along environmental gradients in seasonally dry and fire-prone

ecosystem. Journal of Vegetation Science 22, 1009-1020.

Phillips, R.D., Hopper, S.D., Dixon, K.W., 2010. Pollination ecology and the possible

impacts of environmental change in the southwest Australian biodiversity hotspot.

Philosophical Transactions of the Royal Society B: Biological Sciences 365, 517-528.

Pickett, S.T.A., 1990. Space-for-time substitution as an alternative to long-term studies.

In: Likens, G.E. (Ed.), Long-term studies in Ecology: approaches and alternatives.

Springer-Verlag, New York, pp. 110–135.

Pickett, S.T.A., Burch, W.R., Dalton, S.E., Foresman, T.W., Grove, J.M., Rowtree, R.,

1997. A conceptual framework for the study of human ecosystems in urban areas.

Urban Ecosystems 1, 185-199.

Pickett, S.T.A., Cadenasso, M.L., 2006. Advancing urban ecological studies:

frameworks, concepts, and results from the Baltimore Ecosystem Study. Austral

Ecology 31, 114-125.

Pickett, S.T.A., Cadenasso, M.L., Grove, J.M., 2005. Biocomplexity in coupled natural–

human systems: a multidimensional framework. Ecosystems 8, 225-232.

Pickett, S.T.A., Cadenasso, M.L., Grove, J.M., Groffman, P.M., E. Band, L., Boone,

C.G., Burch Jr., W.R., Grimmond, C.S.B., Hom, J., Jenkins, J.C., Law, N.L., Nilon,

C.H., Pouyat, R.V., Szlavecz, K., Warren, P.S., Wilson, M.A., 2008. Beyond urban

128

legends: an emerging framework of urban ecology, as illustrated by the Baltimore

Ecosystem Study. BioScience 58, 139-150.

Pickett, S.T.A., Cadenasso, M.L., Grove, J.M., Nilon, C.H., Pouyat, R.V., Zipperer,

W.C., Costanza, R., 2001. Urban ecological systems: linking terrestrial, ecological,

physical, and socioeconomic components of metropolitan areas. Annual Review of

Ecology and Systematics 32, 127-157.

Pouyat, R., Yesilonis, I., Szlavecz, K., Csuzdi, C., Hornung, E., Korsós, Z., Russell-

Anelli, J., Giorgio, V., 2008. Response of forest soil properties to urbanization gradients

in three metropolitan areas. Landscape Ecology 23, 1187-1203.

Pouyat, R.V., McDonnell, M.J., Pickett, S.T.A., 1997. Litter decomposition and

nitrogen mineralization in oak stands along an urban-rural land use gradient. Urban

Ecosystems, 117-131.

Pouyat, R.V., Turechek, W.W., 2001. Short- and long-term effects of site factors on net

N-mineralization and nitrification rates along an urban-rural gradient. Urban

Ecosystems 5, 159-178.

Pressey, R.L., Cabeza, M., Watts, M.E., Cowling, R.M., Wilson, K.A., 2007.

Conservation planning in a changing world. Trends in Ecology and Evolution 22, 583-

592.

Pyšek, P., Jarošík, V., Hulme, P.E., Kühn, I., Wild, J., Arianoutsou, M., Bacher, S.,

Chiron, F., Didžiulis, V., Essl, F., Genovesi, P., Gherardi, F., Hejda, M., Kark, S.,

Lambdon, P.W., Desprez-Loustau, M.-L., Nentwig, W., Pergl, J., Poboljšaj, K.,

Rabitsch, W., Roques, A., Roy, D.B., Shirley, S., Solarz, W., Vilà, M., Winter, M.,

2010. Disentangling the role of environmental and human pressures on biological

invasions across Europe. Proceedings of the National Academy of Sciences 107, 12157-

12162.

129

Qian, S.S., Cuffney, T.F., Alameddine, I., McMahon, G., Reckhow, K.H., 2010. On the

application of multilevel modeling in environmental and ecological studies. Ecology 91,

355-361.

Quero, J., Sterck, F., Martínez-Vilalta, J., Villar, R., 2011. Water-use strategies of six

co-existing Mediterranean woody species during a summer drought. Oecologia 166, 45-

57.

R Development Core Team, 2011. R: a language and environment for statistical

computing. R Foundation for Statistical Computing, Vienna, Austria.

Radeloff, V.C., Stewart, S.I., Hawbaker, T.J., Gimmi, U., Pidgeon, A.M., Flather, C.H.,

Hammer, R.B., Helmers, D.P., 2010. Housing growth in and near United States

protected areas limits their conservation value. Proceedings of the National Academy of

Sciences 107, 940-945.

Reckien, D., Martinez-Fernandez, C., 2011. Why do cities shrink? European Planning

Studies 19, 1375-1397.

Regan, H.M., Crookston, J.B., Swab, R., Franklin, J., Lawson, D.M., 2010. Habitat

fragmentation and altered fire regime create trade-offs for an obligate seeding shrub.

Ecology 91, 1114-1123.

Robinson, L., Newell, J.P., Marzluff, J.M., 2005. Twenty-five years of sprawl in the

Seattle region: growth management responses and implications for conservation.

Landscape and Urban Planning 71, 51-72.

Roche, S., Koch, J.M., Dixon, K.W., 1997. Smoke enhanced seed germination for mine

rehabilitation in the southwest of Western Australia. Restoration Ecology 5, 191-203.

Rodewald, A.D., Kearns, L.J., Shustack, D.P., 2011. Anthropogenic resource subsidies

decouple predator–prey relationships. Ecological Applications 21, 936-943.

130

Ross, K.A., Fox, B.J., Fox, M.D., 2002. Changes to plant species richness in forest

fragments: fragment age, disturbance and fire history may be as important as area.

Journal of Biogeography 29, 749-765.

Royal Botanic Gardens Kew, 2008. Seed Information Database (SID) Version 7.1.

Sanderson, E.W., Huron, A., 2011. Conservation in the city. Conservation Biology 25,

421-423.

Sattler, T., Borcard, D., Arlettaz, R., Bontadina, F., Legendre, P., Obrist, M.K., Moretti,

M., 2010. Spider, bee, and bird communities in cities are shaped by environmental

control and high stochasticity. Ecology 91, 3343-3353.

Saunders, D.A., Hobbs, R.J., Margules, C.R., 1991. Biological consequences of

ecosystem fragmentation: a review. Conservation Biology 5, 18-32

Schleicher, A., Biedermann, R., Kleyer, M., 2011. Dispersal traits determine plant

response to habitat connectivity in an urban landscape. Landscape Ecology 26, 529-540.

Schlesinger, M.D., Manley, P.N., Holyoak, M., 2008. Distinguishing stressors acting on

land bird communities in an urbanizing environment. Ecology 89, 2302-2314.

Schneider, A., Woodcock, C.E., 2008. Compact, dispersed, fragmented, extensive? A

comparison of urban growth in twenty-five global cities using remotely sensed data,

pattern metrics and census information. Urban Studies 45, 659-692.

Seddon, G., 1972. Sense of place: a response to an environment, the Swan coastal plain,

Western Australia. University of Western Australia Press, Nedlands, Australia.

Semeniuk, V., Glassford, D.K., 1989. Bassendean and Spearwood Dunes: their

geomorphology, stratigraphy and soils as a basis for habitats of Banksia woodlands.

Journal of the Royal Society of Western Australia 71, 87–88.

Seto, K.C., Fragkias, M., Güneralp, B., Reilly, M.K., 2011. A meta-analysis of global

urban land expansion. PLoS ONE 6, e23777.

131

Seto, K.C., Kaufmann, R.K., Woodcock, C.E., 2000. Landsat reveals China's farmland

reserves, but they're vanishing fast. Nature 406, 121-121.

Seto, K.C., Sánchez-Rodríguez, R., Fragkias, M., 2010. The new geography of

contemporary urbanization and the environment. Annual Review of Environment and

Resources 35, 167-194.

Shearer, B.L., Dillon, M., 1996. Susceptibility of plant species in Banksia woodlands on

the Swan Coastal Plain, Western Australia, to infection by Phytophthora cinnamomi.


Shochat, E., Lerman, S.B., Anderies, J.M., Warren, P.S., Faeth, S.H., Nilon, C.H., 2010.

Invasion, competition, and biodiversity loss in urban ecosystems. BioScience 60, 199-

208.

Shochat, E., Warren, P.S., Faeth, S.H., McIntyre, N.E., Hope, D., 2006. From patterns

to emerging processes in mechanistic urban ecology. Trends in Ecology and Evolution

21, 186-191.

Simard, M., Romme, W.H., Griffin, J.M., Turner, M.G., 2011. Do mountain pine beetle

outbreaks change the probability of active crown fire in lodgepole pine forests?

Ecological Monographs 81, 3-24.

Sommer, B., Froend, R., 2011. Resilience of phreatophytic vegetation to groundwater

drawdown: is recovery possible under a drying climate? Ecohydrology 4, 67-82.

Stenhouse, R.N., 2004. Fragmentation and internal disturbance of native vegetation

reserves in the Perth metropolitan area, Western Australia. Landscape and Urban

Planning 68, 389-401.

Stenhouse, R.N., 2005. Assessing disturbance and vegetation condition in urban

bushlands. Australian Journal of Environmental Management 12, 16-26.

Strayer, D.L., Eviner, V.T., Jeschke, J.M., Pace, M.L., 2006. Understanding the long-

term effects of species invasions. Trends in Ecology and Evolution 21, 645-651.

132

Suding, K.N., Lavorel, S., Chapin III, F.S., Cornelissen, J.H.C., Diaz, S., Garnier, E.,

Goldberg, D., Hooper, D.U., Jackson, S.T., Navas, M.-L., 2008. Scaling environmental

change through the community-level: a trait-based response-and-effect framework for

plants. Global Change Biology 14, 1125-1140.

Syphard, A.D., Radeloff, V.C., Hawbaker, T.J., Stewart, S.I., 2009. Conservation

threats due to human-caused increases in fire frequency in Mediterranean-climate

ecosystems. Conservation Biology 23, 758-769.

Tavernia, B.G., Reed, J.M., 2009. Spatial extent and habitat context influence the nature

and strength of relationships between urbanization measures. Landscape and Urban

Planning 92, 47-52.

Theobald, D.M., 2004. Placing exurban land-use change in a human modification

framework. Frontiers in Ecology and the Environment 2, 139-144.

Theoharides, K.A., Dukes, J.S., 2007. Plant invasion across space and time: factors

affecting nonindigenous species success during four stages of invasion. New

Phytologist 176, 256-273.

Thompson, K., McCarthy, M.A., 2008. Traits of British alien and native urban plants.

Journal of Ecology 96, 853-859.

Tilman, D., May, R.M., Lehman, C.L., Nowak, M.A., 1994. Habitat destruction and the

extinction debt. Nature 371, 65-66.

Tkacz, R., Moody, B., Castillo, J.V., E., F.M., 2008. Forest health conditions of North

America. Environmental Pollution 155, 409-475.

Tress, G., Tress, B., Fry, G., 2004. Clarifying integrative research concepts in landscape

ecology. Landscape Ecology 20, 479-493.

UN, 2010. World Urbanization Prospects. The 2009 Revision. Highlights. United

Nations, New York, USA.

133

UNFPA, 2007. State of the World Population 2007. Unleashing the potential of urban

growth. United Nations Population Fund, New York, USA.

Vallet, J., Daniel, H., Beaujouan, V., Rozé, F., Pavoine, S., 2010. Using biological traits

to assess how urbanization filters plant species of small woodlands. Applied Vegetation

Science 13, 412-424.

Vilà, M., Ibáñez, I., 2011. Plant invasions in the landscape. Landscape Ecology 26, 461-

472.

Webb, C.T., Hoeting, J.A., Ames, G.M., Pyne, M.I., Poff, L., 2010. A structured and

dynamic framework to advance traits-based theory and prediction in ecology. Ecology

Letters 13, 267-283.

Weller, R., 2009. Boomtown 2050 - scenarios for a rapidly growing city. University of

Western Australia Press, Nedlands, Australia.

Western Australian Herbarium, 1998-2011. FloraBase - the Western Australian Flora.

Department of Environment and Conservation. (http://florabase.dec.wa.gov.au/).

Whelan, R.J., Burbidge, A.H., 1980. Flowering phenology, seed set and bird pollination

of five Western Australian Banksia species. Australian Journal of Ecology 5, 1-7.

Williams, N.S.G., Morgan, J.W., McCarthy, M.A., McDonnell, M.J., 2006. Local

extinction of grassland plants: the landscape matrix is more important than patch

attributes. Ecology 87, 3000-3006.

Williams, N.S.G., Morgan, J.W., McDonnell, M.J., McCarthy, M.A., 2005. Plant traits

and local extinctions in natural grasslands along an urban-rural gradient. Journal of

Ecology 93, 1203-1213.

Williams, N.S.G., Schwartz, M.W., Vesk, P.A., McCarthy, M.A., Hahs, A.K.,

Clemants, S.E., Corlett, R.T., Duncan, R.P., Norton, B.A., Thompson, K., McDonnell,

M.J., 2009. A conceptual framework for predicting the effects of urban environments on

floras. Journal of Ecology 97, 4-9.

134

Willis, K.J., Birks, H.J.B., 2006. What is natural? The need for a long-term perspective

in biodiversity conservation. Science 314, 1261-1265.

Winfree, R., Aguilar, R., Vázquez, D.P., LeBuhn, G., Aizen, M.A., 2009. A meta-

analysis of bees' responses to anthropogenic disturbance. Ecology 90, 2068-2076.

Wittemyer, G., Elsen, P., Bean, W.T., Burton, A.C.O., Brashares, J.S., 2008.

Accelerated human population growth at protected area edges. Science 321, 123-126.

Wu, J., David, J.L., 2002. A spatially explicit hierarchical approach to modeling

complex ecological systems: theory and applications. Ecological Modelling 153, 7-26.

Wu, J., Loucks, O.L., 1995. From balance of nature to hierarchical patch dynamics: a

paradigm shift in ecology. The Quarterly Review of Biology 70, 439-466.

Young, R.F., 2009. Interdisciplinary foundations of urban ecology. Urban Ecosystems

12, 311–331.

Zencich, S.J., Froend, R.H., Turner, J.V., Gailitis, V., 2002. Influence of groundwater

depth on the seasonal sources of water accessed by Banksia tree species on a shallow,

sandy coastal aquifer. Oecologia 131, 8-19.

Zhu, W.X., Carreiro, M.M., 2004. Variations of soluble organic nitrogen and microbial

nitrogen in deciduous forest soils along an urban–rural gradient. Soil Biology and

Biochemistry 36, 279-288.

Zipperer, W.C., Wu, J., Pouyat, R.V., Pickett, S.T.A., 2000. The application of

ecological principles to urban and urbanizing landscapes. Ecological Applications 10,

685-688.

Zuur, A.F., Ieno, E.N., Walker, N.J., Saveliev, A.A., Smith, G.M., 2009. Mixed effects

models and extensions in ecology with R. Springer, New York, USA.

135

APPENDICES

Appendix A. Map of study area and identification of the 30 remnants sampled

Appendix B. GIS data collected for the study

Appendix C. Configuration of the sampling plot

Appendix D. Qualitative disturbance assessment used to quantify intensity of human

activities and grazing in the study urban remnants

Appendix E. Principal Component Analysis (PCA) of landscape and local factors

Appendix F. Principal Component Analysis of soil nutrient data

Appendix G. List of plant species selected for functional trait analysis

Appendix H. Summary of the fourth corner-analysis results for the shrub and

herbaceous community

136

Appendix A1. Map of the study area showing the location of the 30 remnant Banksia

woodlands in Perth, Western Australia. Main road network and remnant vegetation are

represented in light gray.

137

Appendix A2. Identification of the 30 remnants sampled in the study

Name Area (ha) Council Ownership

Signal Hill Bushland 2.7 Belmont Local gov.

Queens Park Bushland, Bush Forever site no. 283 8.9 Canning State gov.

Allendale Entrance Bushland 7.7 Cockburn State gov.

Armadale - Warton Rd, Bush Forever site no. 390 70.2 Cockburn Local gov.

Bloodwood Park 1.1 Cockburn Local gov.

Classon Park 2.0 Cockburn Local gov.

Jandakot Rd. Private Property 28.8 Cockburn Private

Wattleup Rd. Private Property 3.5 Cockburn Private

Western Power Private Property 13.7 Cockburn State gov.

Aylesford Reserve 1.1 Gosnells Local gov.

Gay St. Private Property 5.7 Gosnells Private

Dundas Rd. Bushland, BF319 94.6 Kalamunda Private

Anketell Rd. Private Property 7.4 Kwinana Private

Frankland Quarry Bushland 82.0 Kwinana Private

Mortimer Rd. Bushland 43.5 Kwinana Private

Hatfield Bolger Park 1.2 Melville Local gov.

Wireless Hill Park, Bush Forever site no. 336 37.0 Melville Local gov.

Goss Av. Reserve 1.4 South Perth Local gov.

Avocado Bushland 1.0 Stirling Local gov.

Cottonwood Crescent, Bush Forever site no. 43 17.5 Stirling Private

Reid Highway Bushland, Bush Forever site no. 385 73.2 Stirling State gov.

Wordsworth Reserve (Inglewood Golf Course) 2.7 Stirling Private

Lightning Swamp Bush Forever site no. 307 72.4 Swan State gov.

Kensington Park, Bush Forever site no. 48 9.5 Victoria Park Local gov.

Badgup Rd. Private Property 4.2 Wanneroo Private

Gnangara Rd. Private Property 7.9 Wanneroo Private

Hawkins Rd. Bushland, Bush Forever site no. 326 313.9 Wanneroo State gov.

Highview Bushland, Bush Forever site no. 493 12.8 Wanneroo State gov.

Koondoola Regional Park, Bush Forever site no. 201 131.1 Wanneroo State gov.

Montrose Park 5.6 Wanneroo Local gov.

138

Appendix B. GIS data collected for the study

Geographic data from the study area was assembled from different institutions and

governmental departments in an ArcGIS 9.3 database (ESRI, 2008). This contained the

following imagery: digitized geo-referenced black-and-white aerial photographs from

1965, 1972, 1981, 1985, and 1991; true-colour orthophotos from 2000 and 2006;

orthorectified Landsat imagery in intervals of 3 years between 1972 and 1988, 2 years

between 1988 and 2002, and every year from 2002 onwards. The aerial photographs and

orthophotos were provided by the Western Australian Land Information Authority

(Landgate) and the satellite imagery by the Department of Environment and

Conservation. The geographic database also included several thematic shape files

mapping the remnant vegetation extent (Department of Planning and Infrastructure,

2006), vegetation complexes (Western Australian Local Government Association

WALGA, 2005), geomorphology and wetlands location (Department of Environment

and Conservation, 2006), soil (Department of Agriculture and Food, 2007), ground

water levels (Department of Water, 2003), and urban planning (Department of Planning

and Infrastructure, 1965 - 2007).

139

Appendix C. Configuration of the sampling plot

Each sampling plot was composed of nested circular areas where different plant and

environmental data were collected. Circular areas had the following radius: (a) 1.5 m,

(b) 5.5 m, and (c) 11 m.

N

W

b

c

a

140

Appendix D. Qualitative disturbance assessment used to quantify intensity of human

activities and grazing in the study remnants

The intensity of human activities (i.e., trampling, waste disposal, and soil physical

disturbance) and grazing (native and non-native herbivores) in the remnants was

systematically assessed using a semi-qualitative disturbance assessment. This was

composed of a series of questions, which were scored from 0 to 5 in ascending order of

significance in the site (0 = absent, non-significant; 1 = very low; 2 = low but

significant; 3 = intermediate; 4 = high; 5 = very high). Six final composite variables,

measuring the intensity of trampling, waste disposal, soil physical disturbance, overall

human activities, grazing by native and non-native herbivores, were calculated by

summing the different scores and dividing by the total maximum value.

Human activities

Trampling: a) Are there diffuse, narrow, vegetated "goat" tracks? b) Are there narrow

cleared tracks?

Waste disposal: a) Are there signs that the plot was previously used for dumping

(landfill)? b) Signs that the site is currently used for dumping? c) Is there presence of

old infrastructures and construction material (e.g., bricks and metal scraps)? d) Rubbish

(e.g., plastic bags and drink bottles)? e) Vegetative rubbish (e.g., palm fronds)?

Soil physical disturbance: a) Are there signs of previous excavations or alteration of the

soil surface (e.g., mounds of dumped sand)? b) Does the area look like it was cleared

before (e.g., vegetation altered, irregular soil surface)? c) Are there cleared areas with

signs of recent disturbance? d) Are there motor ramps and other signs of motorbike

activity? e) Old 4WD tracks?

141

Grazing

Grazing intensity by the Western grey Kangaroo (Macropus fuliginosus): a) Are there

tracks and resting areas? b) Excrement?

Grazing intensity by non-native European rabbit (Oryctolagus cuniculus): a) Are there

burrows? b) Soil diggings and scratches? c) Excrement?

Grazing intensity was also assessed by counting the number of excrements of each

species in the four 1.5 m radius circles and averaging data to the plot. As the composite

and counting variables were highly correlated and the first performed better during

preliminary data analysis, only those were used on remaining analysis.

142

Appendix E. Principal Component Analysis (PCA) of the landscape and local factors

The first three axes of the PCA conducted with all landscape and local factors explained

58% of the total variation. The first principal component PC1 represented a gradient of

time since urbanization (r=-0.85) and, to a less extent, time since isolation (r=-0.76).

Remnants isolated for longer in the urban matrix were originally fragmented for

agricultural development (r=-0.70), had small past (r=0.85) and current (r=0.55) area.

Importantly, these remnants had higher soil fertility, with higher levels of P (r=-0.68),

percent organic C (r=-0.61), and K (r=-0.58). Furthermore, they had higher intensity of

disturbance by trampling (r=0.60). The PC1 was also correlated with a gradient of

rurality (r=0.60), which was also defined by current landscape connectivity (r=0.76)

and distance to the CBD (r=0.71). Rural remnants had higher grazing intensity by non-

143

native herbivores (r=0.50). The PC2 distinguished the remnants that were fragmented

directly by urbanization (r=0.61) from those currently located in the rural matrix (r=-

0.56). Remnant directly fragmented for urbanization had higher past landscape

connectivity (r=0.72), higher fire frequency (r=0.71), lower litter cover (r=-0.67), and

lower levels of soil S (r=-0.67).

144

Appendix F. Principal Component Analysis of soil nutrient data

The first axis of a PCA conducted with the soil data (a) represented a gradient of soil

fertility, being largely explained by the concentration of K (r=-0.85), P (r=-0.75), NO3

(r=-0.67), and percent organic C (r=-0.78). The second axis distinguished remnants

across a gradient of NH4 (r=-0.89) and pH (r=0.86), which might be associated with a

gradient of natural variation between the two main soil types in the area, Spearwood and

Basseandean (b).

145

Appendix G. List of plant species selected for functional trait analysis

* Species with ≥1% abundance <5% and present in more than 25% of the sites. Native

species are identified with N and exotic species are identified with E

Family Genus Species Code Origin

Overstorey

Casuarinaceae Allocasuarina fraseriana AlloFras N

Myrtaceae Corymbia calophylla CoryCalo N

Myrtaceae Eucalyptus marginata EucaMarg N

Myrtaceae Eucalyptus todtiana EucaTodt N

Myrtaceae Melaleuca preissiana MelaPrei N

Proteaceae Banksia attenuata BankAtte N

Proteaceae Banksia ilicifolia BankIlic N

Proteaceae Banksia menziesii BankMenz N

Loranthaceae Nuytsia floribunda NuytFlor N

Proteaceae Persoonia elliptica PersElli N

Understorey

Shrubs

Casuarinaceae Allocasuarina humilis AlloHumi N

Dilleniaceae Hibbertia aurea HibbAure N

Dilleniaceae Hibbertia huegelii HibbHueg* N

Dilleniaceae Hibbertia hypericoides HibbHype N

Dilleniaceae Hibbertia racemosa HibbRace* N

Dilleniaceae Hibbertia subvaginata HibbSubv N

Epacridaceae Astroloma xerophyllum AstrXero N

Epacridaceae Brachyloma preissii BracPrei N

Epacridaceae Conostephium pendulum ConoPend N

Epacridaceae Leucopogon conostephioides LeucCono N

Epacridaceae Leucopogon polymorphus LeucPoly N

Epacridaceae Lysinema ciliatum LysiCili N

Goodeniaceae Lechenaultia floribunda LechFlor N

Goodeniaceae Scaevola repens ScaeRepe N

Lamiaceae Hemiandra linearis HemiLine N

Lamiaceae Hemiandra pungens HemiPung N

Mimosaceae Acacia iteaphylla AcacItea E

Mimosaceae Acacia pulchella AcacPulc N

Myrtaceae Calothamnus quadrifidus CaloQuad N

Myrtaceae Calothamnus sanguineus CaloSang N

Myrtaceae Calytrix angulata CalyAngu N

Myrtaceae Calytrix flavescens CalyFlav N

Myrtaceae Calytrix fraseri CalyFras N

Myrtaceae Eremaea pauciflora EremPauc N

Myrtaceae Hypocalymma robustum HypoRobu N

Myrtaceae Melaleuca scabra MelaScab N

Myrtaceae Melaleuca thymoides MelaThym N

Myrtaceae Regelia ciliata RegeCili N

Myrtaceae Scholtzia involucrata SchoInvo N

Papilionaceae Bossiaea eriocarpa BossErio N

146


Papilionaceae Daviesia divaricata DaviDiva N

Papilionaceae Daviesia nudiflora DaviNudi N

Papilionaceae Daviesia physodes DaviPhys N

Papilionaceae Daviesia triflora DaviTrif N

Papilionaceae Gastrolobium capitatum GastCapi N

Papilionaceae Gompholobium tomentosum GompTome N

Papilionaceae Hardenbergia comptoniana HardComp N

Papilionaceae Jacksonia floribunda JackFlor N

Papilionaceae Jacksonia furcellata JackFurc N

Papilionaceae Jacksonia sericea JackSeri N

Papilionaceae Jacksonia sternbergiana JackSter N

Phyllanthaceae Phyllanthus calycinus PhylCaly N

Pittosporaceae Billardiera fraseri BillFras N

Proteaceae Adenanthos cygnorum AdenCygn N

Proteaceae Banksia dallanneyi BankDall N

Proteaceae Conospermum stoechadis ConoStoe N

Proteaceae Hakea ruscifolia HakeRusc N

Proteaceae Lambertia multiflora var. darlingensis LambMult N

Proteaceae Persoonia saccata PersSacc N

Proteaceae Petrophile linearis PetrLine N

Proteaceae Petrophile macrostachya PetrMacr N

Proteaceae Stirlingia latifolia StirLati N

Rutaceae Philotheca spicata PhilSpic N

Thymelaeaceae Pimelea sulphurea PimeSulp* N

Xanthorrhoeaceae Xanthorrhoea brunonis XantBrun N

Xanthorrhoeaceae Xanthorrhoea preissii XantPrei N

Zamiaceae Macrozamia fraseri MacrFras N

Herbaceous

Anthericaceae Chamaescilla corymbosa ChamCory N

Anthericaceae Laxmannia squarrosa LaxmSqua* N

Anthericaceae Tricoryne elatior TricElat N

Apiaceae Trachymene pilosa TracPilo* N

Asteraceae Hypochaeris glabra HypoGlab E

Asteraceae Ursinia anthemoides UrsiAnth E

Colchicaceae Burchardia congesta BuchCong* N

Cyperaceae Lepidosperma angustatum LepiAngu N

Cyperaceae Lepidosperma scabrum LepiScab* N

Cyperaceae Mesomelaena pseudostygia MesoPseu N

Cyperaceae Schoenus brevisetis SchoBrev N

Cyperaceae Schoenus curvifolius SchoCurv* N

Cyperaceae Tetraria octandra TetrOcta N

Cyperaceae Tricostularia neesii TricNees N

Dasypogonaceae Dasypogon bromeliifolius DasyBrom N

Dasypogonaceae Dasypogon obliquifolius DasiObli N

Dasypogonaceae Lomandra caespitosa LomaCaes* N

Dasypogonaceae Lomandra hermafrodita LomaHerm* N

Dasypogonaceae Lomandra odora LomaOdor* N

Dasypogonaceae Lomandra suaveolens LomaSuav* N

Euphorbiaceae Monotaxis grandiflora MonoGran N

147


Fumariaceae Fumaria capreolata FumaCapr E

Goodeniaceae Dampiera linearis DampLine N

Haemodoraceae Conostylis aculeata ConoAcul N

Haemodoraceae Conostylis aurea ConoAure* N

Haemodoraceae Conostylis setigera ConoSeti* N

Haemodoraceae Haemodorum laxum HaemLaxu* N

Haemodoraceae Phlebocarya ciliata PhleCili N

Hemerocallidaceae Arnocrinum preissii ArnoPrei N

Hemerocallidaceae Corynotheca micrantha CoryMicr N

Hemerocallidaceae Hensmania turbinata HensTurb N

Iridaceae Freesia alba x leichtlinii FreeAlba E

Iridaceae Gladiolus caryophyllaceus GladCary* E

Iridaceae Patersonia occidentalis PateOcci N

Iridaceae Romulea rosea RomuRose E

Iridaceae Watsonia meriana WatsMeri E

Lauraceae Cassytha flava CassFlav N

Loganiaceae Phyllangium paradoxum PhyllPara* N

Poaceae Aira caryophyllea AiraCary* E

Poaceae Amphipogon turbinatus AmphTurb N

Poaceae Austrostipa compressa AustComp* N

Poaceae Briza maxima BrizMaxi E

Poaceae Ehrharta calycina EhrhCaly E

Poaceae Neurachne alopecuroidea NeurAlop N

Poaceae Pentaschistis airoides PentAiro E

Poaceae Vulpia bromoides VulpBrom E

Restionaceae Alexgeorgea nitens AlexNite N

Restionaceae Caustis dioica CausDioc N

Restionaceae Desmocladus fasciculatus DesmFasc N

Restionaceae Desmocladus flexuosus DesmFlex N

Restionaceae Hypolaena exsulca HypoExsu* N

Restionaceae Lepidobolus preissianus LepiPrei N

Restionaceae Lyginia barbata LygiBarb N

Restionaceae Lyginia imberbis LygiImbe N

Stylidiaceae Levenhookia stipitata LeveStip* N

Stylidiaceae Stylidium calcaratum StylCalc* N

Stylidiaceae Stylidium repens StylRepe N

148

Appendix H. Summary of the fourth-corner analysis results for the shrubs and

herbaceous communities. Significant relationships between each functional trait state

and the landscape and local factors across the 30 remnant Banksia woodland are shown.

Positive relationships are indicated with a + sign and negative relationships are

indicated with a – sign. Relationships with p<0.05 are filled in yellow and relationships

with p<0.01 are filled in orange.

149

Landscape fragmentation dynamics Disturbance regimes &

environmental conditions

Tim

e si

nce

urb

aniz

atio

n

Tim

e si

nce

iso

lati

on

Pas

t re

mnan

t ar

ea

Pas

t la

ndsc

ape

connec

tivit

y

Curr

ent

rem

nan

t ar

ea

Curr

ent

landsc

ape

connec

tivit

y

Tra

ject

ory

rura

l-urb

an

Tra

ject

ory

woodla

nd

-urb

an

Tra

ject

ory

rura

l-ru

ral

Gra

zing i

nte

nsi

ty b

y e

xoti

c

her

biv

ore

s

Fir

e fr

equen

cy

Tra

mpli

ng

Soil

fer

tili

ty

Lit

ter

dep

th

Seed dispersal

Shrubs

Unassisted dispersal +

Wind dispersal + +

Internal animal transport - - - -

External animal transport + - - - + - + +

Herbaceous

Unassisted dispersal + + - + - - +

Wind dispersal - - + - + + -


Shrubs

Insect-pollinated - +

Wind-pollinated + +

Bird-pollinated +

Auto-pollinated + + - - +

Herbaceous

Insect-pollinated + +

Wind-pollinated

Auto-pollinated - + +


Shrubs

Arbuscular mycorrhizal +

Root cluster + + - - - + - - + +

N2-fixing + + + - - + - - + +

Ericoid mycorrhizal - - - -

Ectomycorrhizal + +

Non-mycorrhizal - +

Herbaceous

Arbuscular mycorrhizal - - + + - + + - -

Root cluster + + - - + -

Non-mycorrhizal + - + - +

Regeneration strategy

Shrubs

Resprouter + + + - + -

Obligate seeder - - - + - +

Herbaceous

Resprouter + + - + - +

Obligate seeder - - + - + -

this thesis is presented for the degree of doctor of

Documents