An Indicator-Based Model for Assessing Sustainability Performance of Urban Green Infrastructure

CITIES WITHIN GARDENS: An indicator-based model for assessing sustainability performance of urban green infrastructure

Parisa Pakzad

A Thesis submitted for the degree of Doctor of philosophy

Australian Graduate School of Urbanism Faculty of Built Environment The University of New South Wales

September 2017

THE UNIVERSITY OF NEW SOUTH WALES Thesis/Dissertation Sheet Surname or Family name: Pakzad Parisa First name:

Abbreviation for degree as given in the PhD University calendar: Australian Graduate School Faculty: Built Environment School: of Urbanism

Title: Cities within gardens: An indicator-based model for assessing sustainability performance of urban green infrastructure

Abstract: In recent years, the planning, design and installation of “green infrastructure” at the local and city level has been identified as a best practice and nature-based solution to achieving greater urban sustainability and resilience. It is a component of the international urban movement “Smart Cities”. Green infrastructure is an integrated multi-scale network of green spaces within, beyond, and around a city. It provides many benefits - most importantly ecosystem services for human and environmental health. This study aims to develop an indicator-based model using a mixed-method approach as a means to evaluate the performance of urban green infrastructure. This model is composed of a set of sixteen key indicators within four subcategories: ecological; health and well-being; sociocultural; and economic. Each represents key interactions between human health, ecosystem services and ecosystem health. The proposed performance indicators are based on the incorporation of results in three systematic, mixed-method approaches that consist of the development of the Drivers- Pressure-State-Impact-Responses (DPSIR) model specific to this research problem. The DPSIR model is a conceptual foundation to govern the development of sustainability indicators. Semi-structured interviews are undertaken involving twenty-one selected Australian experts, and input from 373 Australian national and international stakeholders from representative fields via an online questionnaire. An assessment matrix is developed that comprises description, calculation (equation) and units for each individual indicator. This model is tested, validated and verified through a case study in Sydney, Australia. The significance of the research is that: the proposed indicator-based model provides an opportunity to understand the complex relationships of the multidimensional structure of urban green spaces; it serves as a useful insight for urban designers and decision-makers in monitoring various aspects of the urban ecosystem; and it also allows for early warnings regarding any undesirable changes in sustainability levels.

Declaration relating to disposition of project thesis/dissertation I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstracts International (this is applicable to doctoral theses only).

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

FOR OFFICE USE ONLY Date of completion of requirements for

Award:

ORIGINALITY STATEMENT

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

Signed …………………………………………….... Date ……………………………………………......

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

Signed …………………………………………….... Date ……………………………………………......

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

Signed …………………………………………….... Date ……………………………………………......

ABSTRACT

In recent years, the planning, design and installation of “green infrastructure” at the local and city level has been identified as a best practice and nature-based solution to achieving greater urban sustainability and resilience. It is a component of the international urban movement “Smart Cities”. Green infrastructure is an integrated multi-scale network of green spaces within, beyond, and around a city. It provides many benefits - most importantly ecosystem services for human and environmental health.

This study aims to develop an indicator-based model using a mixed-method approach as a means to evaluate the performance of urban green infrastructure. This model is composed of a set of sixteen key indicators within four subcategories: ecological; health and well-being; sociocultural; and economic. Each represents key interactions between human health, ecosystem services and ecosystem health. The proposed performance indicators are based on the incorporation of results in three systematic, mixed-method approaches that consist of the development of the Drivers-Pressure-State-Impact-Responses (DPSIR) model specific to this research problem. The DPSIR model is a conceptual foundation to govern the development of sustainability indicators. Semi-structured interviews are undertaken involving twenty-one selected Australian experts, and input from 373 Australian national and international stakeholders from representative fields via an online questionnaire. An assessment matrix is developed that comprises description, calculation (equation) and units for each individual indicator. This model is tested, validated and verified through a case study in Sydney, Australia.

The significance of the research is that: the proposed indicator-based model provides an opportunity to understand the complex relationships of the multidimensional structure of urban green spaces; it serves as a useful insight for urban designers and decision-makers in monitoring various aspects of the urban ecosystem; and it also allows for early warnings regarding any undesirable changes in sustainability levels.

Key words: Green infrastructure performance model, indicators, assessment matrix, sustainability assessment, ecosystem services, ecosystem and human health.

ACKNOWLEDGMENTS

Special thanks to Dr Paul Osmond, Linda Corkery and Dr Graciela Metternicht for their exemplary supervision, who provided vital support and guidance throughout the duration of the research project.

I would like to thank the contribution of the expert participants in the semi-structured interviews and online questionnaire and also the Australian Institute of Landscape Architects, Australian Institute of Architects (AIA), Low Carbon Living CRC, Infrastructure Sustainability Council of Australia (ISCA) and Australian Sustainable Built Environment Council (ASBEC) for helping me to distribute the questionnaire through their professional networks.

I would like to acknowledge the financial support of the project partners (University of New South Wales, the Cooperative Research Centre for Low Carbon Living (CRC-LCL) and Infrastructure Sustainability Council of Australia (ISCA)), Special thanks to Professor Deo Prasad at the CRC Low Carbon Living because this research would not have been possible without his financial and intellectual support.

TABLE OF CONTENTS

ABSTRACT ...... 4 ACKNOWLEDGMENTS ...... 5 TABLE OF CONTENTS ...... 6 LIST OF TABLES ...... 8 LIST OF FIGURES ...... 11 CHAPTER I: INTRODUCTION ...... 13 Introduction ...... 13 1.1 Background information and research problem ...... 16 1.2 Research objectives and questions ...... 17 1.3 Methodological framework ...... 18 1.4 Research contributions ...... 22 1.5 Thesis outline ...... 23 Publications ...... 25 CHAPTER 2: LITERATURE REVIEW ...... 26 Introduction ...... 26 Green Infrastructure (GI) ...... 26 2.1 Green infrastructure concept and terminology ...... 26 2.2 History of green infrastructure thinking ...... 30 2.3 Green infrastructure components and key principles ...... 37 2.4 Review of existing GI assessment models and applications ...... 40 2.4.1 Detailed analysis of three green infrastructure evaluation tools ...... 48 2.5 Green infrastructure conceptual frameworks in the literature ...... 54 2.5.1 Existing frameworks for assessing urban sustainability ...... 56 2.5.2 Existing green infrastructure conceptual models ...... 58 2.6 Redefining the GI conceptual framework ...... 60 2.6.1 Ecosystem services parameters ...... 61 2.6.2 Human health parameters ...... 65 2.6.3 Ecosystem health parameters ...... 70 2.7 An integrated approach: Linking GI with ecosystem services, human health and ecosystem health ...... 71 Summary ...... 73 CHAPTER 3: METHODOLOGY ...... 74 Introduction ...... 74 3.1 Research design ...... 74 3.2 Mixed-method approach ...... 75 3.2.1 Qualitative-method approach ...... 76 3.2.2 Quantitative-method approach ...... 78 Summary ...... 78 CHAPTER 4: INTERVIEW ANALYSIS ...... 80 Introduction ...... 80 4.1 Semi-structured research interview design and data collection approach ...... 81 4.2 Data analysis ...... 81 Section 1: Familiarity with existing rating tools ...... 82 Section 2: Understanding of the green infrastructure concept ...... 82 Section 3: Establishing the framework ...... 86 4.3 Stakeholders’ mind mapping ...... 86 4.4 Proposed green infrastructure performance indicators (Initial list) ...... 88

Summary ...... 90 CHAPTER 5: ONLINE QUESTIONNAIRE ANALYSIS ...... 91 Introduction ...... 91 5.1 Questionnaire design ...... 92 5.1.1 Piloting ...... 94 5.1.2 Sampling ...... 94 5.2 Questionnaire findings ...... 96 5.2.1 Participants profile (Results from part C of the questionnaire) ...... 96 5.2.2 Results from section A of the questionnaire ...... 99 5.2.3 Results from section B of the questionnaire ...... 104 5.3 Analysis to identify the important indicators in measuring the level of sustainability 105 5.4 Determining key indicators ...... 108 Summary ...... 108 CHAPTER 6: DEVELOPING GIS ASSESSMENT METRICS ...... 110 Introduction ...... 110 6.1 Description of selected indicators ...... 110 6.2 Ecological indicators ...... 111 6.3 Health indicators ...... 157 6.4 Sociocultural indicators ...... 166 6.5 Economic indicators ...... 173 Summary ...... 176 CHAPTER 7: Case study application of the model ...... 177 Introduction to the case study ...... 177 7.1 General characteristics of the study area ...... 177 7.2 Green structure of the study area ...... 185 7.3 Data acquisition and preparation ...... 186 7.4 Methodology of data analysis ...... 187 7.5 Review of results...... 194 7.6 Overall results ...... 220 Summary ...... 223 CHAPTER 8: CONCLUSIONS ...... 224 Introduction ...... 224 8.1 Discussion of findings in relation to the research objectives ...... 225 8.1.1 Establishment of the conceptual framework ...... 226 8.1.2 Investigation of key indicators ...... 227 8.1.3 Development of metrics for indicator-based model ...... 228 8.1.4 Validation of the proposed model via a case study ...... 229 8.2 Significance of the research ...... 229 8.3 Limitations and recommendations for future research ...... 230 REFERENCES ...... 232 APPENDIX A: Online questionnaire form...... 257 SECTION A: Verify green infrastructure performance framework ...... 257 SECTION B: Rating indicators ...... 258 SECTION C: Expert classification ...... 260 APPENDIX B: Summary of questionnaire results ...... 262 APPENDIX C: Frequency of species and calculation of Shannon diversity score ...... 268

LIST OF TABLES

Table Page Table 2. 1 Definitions of green infrastructure in various disciplines ...... 27 Table 2. 2 Key benefits of green infrastructure in literature ...... 29 Table 2. 3 Green infrastructure movement milestones ...... 30 Table 2. 4 Green infrastructure key principles (Mell 2008, 73; revised by author) ...... 38 Table 2. 5 Green infrastructure evaluation tools ...... 41 Table 2. 6 Intersection between the GI features and ecosystem services and biodiversity(Source: Author)...... 53 Table 2. 7 Models and theories linking ecosystem and human health aspects ...... 59 Table 2. 8 Millennium Ecosystem Services Framework (Millennium Ecosystem Assessment 2003, 57) ...... 61 Table 2. 9 Ecosystem services that can be provided by green infrastructure ...... 63 Table 2. 10 GI contributes to ecosystem and human health through services delivered by ecosystem (Noss and Cooperrider 1994; Tzoulas et al. 2007 and Austin 2014; revised by author)...... 64 Table 2. 11 Studies which have defined ecosystem health ...... 70

Table 3. 1 Strengths and weaknesses of mixed method approach ...... 76 Table 3. 2 Strengths and weaknesses of qualitative and quantitative research methods ...... 77 Table 3. 3 The characteristics of interview types ...... 77 Table 3. 4 Research Objectives, Questions and Approaches ...... 79

Table 4. 1 Interviewee profiles ...... 80 Table 4. 2 Descriptive statistics of answers for Yes/No questions ...... 81 Table 4. 3 Identifying the most popular tools ...... 82 Table 4. 4 GI definitions among three groups of participants ...... 83 Table 4. 5 The most well-known GI’s components ...... 83 Table 4. 6 Weighting of GI benefits ...... 85 Table 4. 7 GI structure ...... 85 Table 4. 8 Benchmarking scale ...... 86 Table 4. 9 GI thematic concepts derived from the interview results ...... 88 Table 4. 10 proposed green infrastructure performance indicator set ...... 89

Table 5. 1 Quantitative Questionnaire Design...... 93 Table 5. 2 Distribution of Australian participants ...... 97 Table 5. 3 Distribution of International participants ...... 97 Table 5. 4 Participants classifications based on job sectors...... 98 Table 5. 5 Disaggregation of participants by industry sector ...... 98 Table 5. 6 Number of years’ work experience ...... 99 Table 5. 7 Section A of the questionnaire ...... 100 Table 5. 8 Descriptive statistics relating to definitions and job sectors ...... 102 Table 5. 9 Weighting sub-categories of the MEA framework ...... 104 Table 5. 10 Weighting sub-categories of the TBL framework ...... 104 Table 5. 11 Three frameworks ...... 105

Table 5. 12 Weighting main categories result ...... 105 Table 5. 13 Indicators included in questionnaire and weighting indicators ...... 106 Table 5. 14 Key indicators set ...... 109

Table 6. 1 Normalisation of indicators by setting baseline values ...... 111 Table 6. 2 Ecological indicators...... 112 Table 6. 3 The approximate ranges of surface albedos ...... 117 Table 6. 4 Normalisation of surface albedo value ...... 117 Table 6. 5 Evapotranspiration Pan Data- Mean daily (mm per day) ...... 120 Table 6. 6 Landscape coefficients range from very low to high ...... 121 Table 6. 7 Categories of water needs ...... 121 Table 6. 8 Estimated values for species and density, microclimate...... 122 Table 6. 9 Normalisation of evapotranspiration rate ...... 123 Table 6. 10 Beaufort scale ...... 125 Table 6. 11 Wind direction in Sydney - Principal times of the year for wind occurrence in Sydney...... 126 Table 6. 12 Normalisation of ventilation value for Sydney ...... 126 Table 6. 13 Normalisation of the shading effect of trees ...... 129 Table 6. 14 Normalisation of air pollutant removal ratio ...... 134 Table 6. 15 Normalisation of carbon sequestration ratio ...... 139 Table 6. 16 Normalization values for energy use reduction - only cooling (mature tree height) ...... 141 Table 6. 17 Event Mean Concentration (EMC) ...... 143 Table 6. 18 Tree canopy relative land use loading rates based on underlying land-use landcover ...... 144 Table 6. 19 Quality of urban runoff in Australia and global ...... 144 Table 6. 20 Normalisation of water quality improvements ratio ...... 145 Table 6. 21 Typical TN, TP and TSS Removal rates, drainage size and requires space for different BMP types ...... 146 Table 6. 22 surface runoff ratio...... 148 Table 6. 23 Normalisation of avoided surface runoff ratio ...... 149 Table 6. 24 Runoff reduction capabilities for various BMPs (Battiata et al. 2010; Hirschman et al. 2008) ...... 149 Table 6. 25 Generic resistance scores of habitats suitability (based on recorded fauna species occurrence data in Parramatta LGA)...... 153 Table 6. 26 Normalisation of habitat connectivity (habitat suitability) value ...... 154 Table 6. 27 Normalisation of species diversity value or Shannon’s equitability score ...... 157 Table 6. 28 Health category ...... 158 Table 6. 29 Normalisation of accessibility value ...... 161 Table 6. 30 Node classifications used in this study ...... 162 Table 6. 31 Normalisation of GI equity index value ...... 163 Table 6. 32 Normalisation of visual accessibility value ...... 165 Table 6. 33 Sociocultural category ...... 166 Table 6. 34 Code SEPP landscaping requirements ...... 167 Table 6. 35 Normalisation of land productivity capacity value ...... 168 Table 6. 36 Normalisation of land use mixed value ...... 169 Table 6. 37 Shared path ...... 171 Table 6. 38 Separate two-way path widths ...... 171 Table 6. 39 Separate one-way path widths ...... 171

Table 6. 40 Normalisation of walkability score ...... 171 Table 6. 41 Normalisation of intersection density value ...... 173 Table 6. 42 Economic category ...... 174

Table 7. 1 Summary of monthly weather statistics in Parramatta ...... 178 Table 7. 2 Monthly prevailing wind wave direction in Sydney (Direct NE, Prevent S and W) ...... 179 Table 7. 3 Percentage of individual land uses in study area (Results derived from GIS data analysis) ...... 182 Table 7. 4 Ten most dominant tree genera in study area...... 185 Table 7. 5 Summary of equations, variables and baseline value level ...... 188 Table 7. 6 Case study - Effective surface albedo ...... 195 Table 7. 7 Case study - Evapotranspiration rate ...... 197 Table 7. 8 Case study – Ventilation ...... 198 Table 7. 9 Case study - Shading effect ...... 199 Table 7. 10 Case study results - Proposed trees with good shade effect performance ...... 199 Table 7. 11 Case study - Air pollutant removal ...... 201 Table 7. 12 Case study - Carbon storage and sequestration ...... 202 Table 7. 13 Case study, annual energy savings due to trees near buildings ...... 203 Table 7. 14 Case study - Energy saving ...... 204 Table 7. 15 Case study - Water quality ...... 205 Table 7. 16 Case study - Avoid surface runoff ...... 206 Table 7. 17 Case study – Habitat connectivity ...... 207 Table 7. 18 Case study – Species diversity ...... 209 Table 7. 19 Case study – Physical accessibility ...... 210 Table 7. 20 Functional classification of GI types and their service radius zone ...... 211 Table 7. 21 Case study – GI equity ...... 213 Table 7. 22 Case study – visual accessibility ...... 214 Table 7. 23 Case study – Food production ...... 215 Table 7. 24 Case study – Mixed-use neighbourhood ...... 216 Table 7. 25 Case study – Walkability ...... 218 Table 7. 26 Case study – Connectivity (intersection density) ...... 219 Table 7. 27 Overall score of existing GI sustainability performance in Parramatta CBD ..... 220 Table 7. 28 Case study – Dollar value of selected economic indicators ...... 223

Table A. 1 Blank questionnaire form ...... 257

Table B. 1 Descriptive statistic of answers for questions 1 to 3 in section A ...... 262 Table B. 2 Descriptive statistic of answers for questions 4 in section A ...... 262 Table B. 3 Descriptive statistic of answers for questions 5 in section A ...... 263 Table B. 4 Weighted average index (WAI)- Ecological indicators ...... 266 Table B. 5 Weighted average index (WAI)- Health indicators ...... 266 Table B. 6 Weighted average index (WAI)- Sociocultural indicators ...... 267 Table B. 7 Weighted average index (WAI)- Economic indicators ...... 267

Table C. 1 Species diversity...... 268

LIST OF FIGURES

Figure Page Figure 1. 1 Interrelationship between ecosystem services, aspects of human well-being and human health ...... 15 Figure 1. 2 Construction of an index ...... 21

Figure 2. 1 A schematic green network in a city ...... 28 Figure 2. 2 History of green infrastructure planning ...... 31 Figure 2. 3 Emerald Necklace map, Franklin Park, Boston, Massachusetts ...... 32 Figure 2. 4 Ebenezer Howard's vision of a garden city and social city (1898)...... 32 Figure 2. 5 The regional reserve system proposed by Larry Harris and Reed Noss ...... 34 Figure 2. 6 Florida GIS-based model: Greenways System Planning Project, 1999 ...... 35 Figure 2. 7 Green infrastructure components ...... 38 Figure 2. 8 Synthesis diagram of green infrastructure ...... 40 Figure 2. 9 Dual rights-based approach ...... 56 Figure 2. 10 DPSIR framework of linkage between human activities and GI performance ... 57 Figure 2. 11 Conceptual framework of green infrastructure performance assessment ...... 60 Figure 2. 12 Cascade model for linking ecosystems to human well-being (Haines-Young and Potschin 2010 and de Groot et al. 2010, 11) ...... 63 Figure 2. 13 Aspects of healthy life ...... 66 Figure 2. 14 Conceptual framework of green infrastructure based on literature review ...... 72

Figure 4. 1 Importance of potential benefits of GI as designated by stakeholders ...... 84 Figure 4. 2 Stakeholders’ mind mapping ...... 87

Figure 5. 1 Participants’ industry sectors ...... 98 Figure 5. 2 Participants’ background and experience ...... 98 Figure 5. 3 Participants' specific field and focus areas ...... 99 Figure 5. 4 Summary of answers given by respondents in Q4 ...... 101 Figure 5. 5 Identifying the appropriate framework referring to question 5 ...... 103

Figure 6. 1 The partitioning of evapotranspiration into evaporation and transpiration over the growing period for an annual field crop ...... 118 Figure 6. 2 Street orientations and H/W ratio (Ali-toudert and Mayer 2006)...... 124 Figure 6. 3 Wind rose- Sydney airport (1995-2014)- all months 10m height- Calm- 0.6% .. 126 Figure 6. 4 Schematic figure of required iTree input data ...... 138 Figure 6. 5 iTree Eco- precipitation interception model diagram ...... 148 Figure 6. 6 Core Reserves, Buffer Zones, and Linkages ...... 155 Figure 6. 7 Explanation of Floor Green View Index (Yu et al. 2016)...... 164 Figure 6. 8 Typical cross-section of a separated protected bicycle lane ...... 170 Figure 6. 9 Comparison between low and high connected network between two points ...... 172

Figure 7. 1 Parramatta local government area (LGA), CBD and green infrastructure distribution...... 178 Figure 7. 2 Annual wind rose for Parramatta ...... 179 Figure 7. 3 Land surface classification of study area (pervious and impervious)...... 180 Figure 7. 4 Local climate zone classifications ...... 181 Figure 7. 5 Typical cross-sectional canyons ...... 181

Figure 7. 6 Mean height-to-width ratio of street canyons classifications...... 182 Figure 7. 7 Land use zoning classifications ...... 183 Figure 7. 8 Distribution of population density (Residents)...... 184 Figure 7. 9 GI distribution (links and nodes) ...... 185 Figure 7. 10 Percentage of tree population by diameter at breast height (DBH)...... 186 Figure 7. 11 Spatial aggrigation of the surface albedo ...... 193 Figure 7. 12 Monthly pollution removal by urban GI in the case study Parramatta CBD .... 200 Figure 7. 13 Annual gross carbon sequestration with existing urban GI in the case study area Parramatta CBD ...... 202 Figure 7. 14 Avoid runoff and value for species with greatest overall impact on runoff ...... 206 Figure 7. 15 Euclidean distance - footpath analysis ...... 210 Figure 7. 16 Travel cost surface analysis ...... 210 Figure 7. 17 Spatial functional distribution of GI types ...... 212 Figure 7. 18 Service radius covered by individual GI types ...... 212 Figure 7. 19 Visible building’s facade analysis ...... 214 Figure 7. 20 Footpath and cycleways ...... 217 Figure 7. 21 Network analysis and interest points (green open spaces)...... 218

Figure 8. 1 Overview of development of the model ...... 225 Figure 8. 2 The DPSIR framework showing the link between human activities and GI performance...... 226

CHAPTER I: INTRODUCTION

Introduction

Rapid urbanisation is a recent phenomenon in the developing world. The United Nations predicts that over the next four decades the urban areas of the world will absorb all the population growth that is expected (United Nations 2012). This will impose a tremendous ecological footprint on local environments, on the planet and on all development scales in between. The rising level of urbanisation has a direct correlation to the increased production and consumption of goods, services and infrastructure. This puts additional pressure on land consumption in the built environment. The resulting problems are landscape fragmentation, biodiversity loss, an increase in the impact of natural hazards (floods, landslides, forest fires, urban heat islands, greenhouse gas emissions) and the destruction of sensitive ecosystems (Heinberg 2010; McKinney 2002; Millennium Ecosystem Assessment 2005b), as well as a decrease in human health and well-being (Coutts 2016; Dale 1997; Tzoulas et al. 2007). In addition, all these above phenomena will interact with, and generally be exacerbated by, climate change (Gill et al. 2007).

Accordingly, over the past few decades, concerns over environmental issues that have affected our planet and the living natural world have resulted in an increasing substantial body of studies focusing on the health of ecosystems, settlements and inhabitants. Guidelines and strategic directions for current and future challenges have been formulated and many guidelines have taken a prime place on the agenda in many cities across the world. Today, there is much more focus on improving the sustainable performance of cities, including buildings and infrastructure, for the benefit of inhabitants and the long-term sustainability of the planet.

In this regard, a number of guidelines, measuring tools and rating applications have been developed to assess the sustainability performance of buildings, their construction processes and that of grey (hard) infrastructure. Some of these rating applications are: LEED (USA); BREEAM (UK); Green Star; NABERS and the Infrastructure Sustainability (IS) rating tools (Australia); and CASBEE (Japan). These tools outline the benefits and outcomes of the decisions made in the creation of the built environment and infrastructure (Vandepol 2010).

Infrastructure systems, both grey and green, are vital to enhancing the functions and services that are provided in modern cities (Neuman 2006; Neuman and Smith 2010). The Oxford English dictionary defines grey infrastructure as “basic physical and organisational structures and facilities which are needed for the operation of a society or enterprise: the social and economic infrastructure of a country” (Ely and Pitman 2012,29) . Traditional grey infrastructure comprises the engineered network of roads and services that deliver a range of goods and services to the population of a city. These infrastructure systems require major capital investment to build and maintain, and are generally single-use occupiers of large areas of urban land (Wolf 2003).

Grey or technical infrastructures facilitate and support social and economic production (Howes and Robinson 2005; Pol 2010). However, due to both their design and operation, they tend to be very costly and can produce large carbon footprints. The outcomes from these types of developments can be harmful to public and individual health (Black and Black 2009; Graham and Marvin 2001); further, they usually serve a single purpose and typically are not integrated well with their environments.

To remedy some of the negative consequences of non-integrated development, green infrastructure - as opposed to grey infrastructure - is identified as a nature-based and cost- effective solution for improving the sustainability of urban development. Sustainability in general refers to the capacity to be continued indefinitely. The most frequently cited definition is that “Development that meets the needs of the present without compromising the ability of future generations to meet their own needs.” (World Commission on Environment and Development 1989). Sustainable development represents a sustainability approach in the context of three pillars or ‘bottom lines’ – ecological, social and economic. Ecologically sustainable development was defined by the Commonwealth of Australia (1992) as “…using, conserving and enhancing the community’s resources so that ecological processes, on which life depends, are maintained, and the total quality of life, now and in the future, can be increased” (Ecologically Sustainable Development Steering Committee 1992). Hence, the ecological perspective of sustainability applied in this study embraces the integration of natural and living systems into urban design with the objective of supporting the well-being and health of society and ecosystems over time.

As such, green infrastructure planning has been identified as the distinctive concept of sustainable urban form and eco-city. It develops over time through evolution at a lower capital, maintenance and operational cost, has fewer negative impacts on the environment and it can significantly reduce the carbon footprint compared to grey infrastructure (Benedict and McMahon 2006; Lafortezza et al. 2013). GI solutions have become increasingly valued for their potential to address a wide variety of services, e.g. water purification and air quality control, as well as to contribute to climate change adaptation and mitigation.

Green infrastructure (GI) has been defined in various ways with the three most common approaches focusing on GI principles of multiple benefits, multi-functionality and connectivity. GI may be described in terms of:

1) the role of ecosystem services as GI’s multiple benefits (Coutts 2016; de Groot et al. 2012; Tzoulas et al. 2007);

2) green engineering as its multiple functions (Margolis and Robinson 2007);

3) linked and integrated green spaces as its connected structure (Benedict and McMahon 2002).

The most valuable definition, however, is considered to be a synthesis of all three (Foster et al. 2011). The Australian Institute of Landscape Architects (2012 ,4) define green infrastructure as:

“… a network of natural landscape assets which underpin the economic, sociocultural and environmental functionality of our cities and towns – i.e. the green spaces and water systems which intersperse, connect, and provide vital life support for humans and other species within our urban environments”.

Pitman et al. (2015) have described GI as the network of green spaces and water bodies that deliver multiple environmental, social and economic benefits and services to urban communities. This network comprises green and blue components such as parks and reserves, gardens and backyards, waterways and wetlands, streets and green transport corridors, pathways and greenways, farms and orchards, squares and plazas, roof gardens and living walls, sports fields and green cemeteries.

GI can provide a wide range of engineering, environmental and human services, collectively known as ecosystem services. Such services (Millennium Ecosystem Assessment 2003) include: provisioning (such as food, water, fiber and fuel); supporting (such as soil formation and nutrient cycling); regulating (such as climate, flood and disease control regulation, and water purification); and cultural services (such as aesthetic, spiritual, symbolic, educational and recreational) , as illustrated by Figure 1.1. Ecosystem services, according to the Millennium Ecosystem Services (Millennium Ecosystem Assessment 2005b), are the missing link between ecosystems and human well-being.

Figure 1. 1 Interrelationship between ecosystem services, aspects of human well-being and human health (Millennium Ecosystem Assessment 2005b, 14)

A further advantage of a green infrastructure network is that it serves multiple functions such as environmental, social, economic and aesthetic functions that are underpinned by healthy ecosystems. In comparison, grey infrastructure tends to be designed to perform only a single function such as moving road traffic. For example, green infrastructure can reduce runoff (ecological); increase community interaction (social); increase physical activities (health); and increase property values (economic) in the neighboring blocks.

Moreover, green infrastructure can be more easily designed as an integrated network where the output of one infrastructure component becomes the input to another. Therefore, the connectedness of the system - whether natural, human-made or a combination of both - is a key positive feature of its sustainability.

1.1 Background information and research problem

In order to address the above challenges, and to promote the adoption of green infrastructure planning, new thinking is needed that considers the fundamental factors and principles of GI: multi-types; multi-scales; multi-functions; and connectivity and integration. The research requirement therefore is a holistic approach embraced in one single model that evaluates the key performances of green infrastructure. Such a holistic approach does not yet exist with available assessment models.

Based on a review of the literature, most of the green infrastructure research studies have taken place in the United States and Europe. However, a number of significant studies have also been undertaken in Australia (City of Melbourne 2011; Fam et al. 2008; GHD 2011; Killicoat et al. 2002; Moore 2000; Norton et al. 2015; Pitman et al. 2015; Plant 2006; Tarran 2006, 2009; Townend and Weerasuriya 2010). Researchers have explored the green infrastructure concept in various disciplines and practices worldwide such as climate change mitigation and adaptation (Australian Institute of Landscape Architect (AILA) 2012; Derkzen et al. 2017; McPherson and Simpson 2009; Moore 2006); urban heat island effect mitigation (Coutts et al. 2007; Di Leo et al. 2016; Livesley 2010; Loughnan et al. 2010; Wang et al. 2015); urban resilience and sustainability (Ahern 2010; Desha et al. 2016); sustainable water management systems and blue/green infrastructure (Wong et al. 2012); human health and well-being (Austin 2014; Coutts 2016; Tzoulas et al. 2007), including recent Australian-based research (Beatley and Newman 2012; Kent et al. 2011; Planet Ark 2012).

As reviewed in the next chapter, there are only a few tools available that assess the performance of green infrastructure, let alone a single integrated tool that can be applied across a wide range of criteria. Moreover, the review shows that there is a lack of consistency amongst the tools available. The lack of a comprehensive tool to assess green infrastructure performance detracts from the value of the tools available. This inhibits the ability to compare the costs and the associated benefits of green versus grey infrastructure. Additionally, unproven green infrastructure approaches create uncertainty, making it difficult to convince decision makers to

16 invest in green infrastructure in order to achieve financial trade-offs since fiscal cost- effectiveness is generally given the priority in formulating businesses cases and in the decision- making process about infrastructure (Talberth et al. 2013).

The predominant theme that emerges from the literature review is that almost all of these studies introduce the benefits and performance of green infrastructure, yet there is no comprehensive, integrated and conclusive way to measure the sustainability performance of green infrastructure. The existing frameworks, tools and methods that measure the performance of green infrastructure are neither comprehensive enough, nor applicable to the wide range of circumstances that this initiative demands. As with other components of the built environment, green infrastructure requires an objective and practical assessment tool to measure its effectiveness, its costs and benefits, and its subsequent overall value. Therefore, a case can be convincingly made to implement these systems more widely with the realisation of benefits. Hence, a comprehensive model is required that could moderate this uncertainty by evaluating the GI benefits and its contribution to reaching sustainability targets in projects.

1.2 Research objectives and questions

To promote and to effectively assess the nature of green infrastructure benefits, a multi-faceted assessment method for measuring its performance and efficiency is essential. Numerous researchers have examined the performance of green infrastructure in reducing the discharge of pollutants into waterways, removing air pollutants from the environment, and even in reducing energy use but there is a lack of empirical multidimensional studies to evaluate the performance of green infrastructure. Assessment and evaluation of green infrastructure is still a developing field that needs more rigour and structure to determine GI’s full and best benefits. The main objective of this research, therefore, is to develop a comprehensive sustainability assessment model for measuring the performance of green infrastructure.

To determine the most effective criteria for such a model, the study will focus on the following specific objectives:

1- Develop a conceptual framework based on a comprehensive literature review and adjust it for the Australian context.

2- Identify and understand the logical and rational relationships between the various indicators of sustainability in relation to green infrastructure performance in order to identify the importance of each indicator.

3- Determine and define appropriate methods for formulating indicators that measure the degree of sustainability performance. This objective will establish a baseline/benchmark level for each indicator.

4- Test the pilot assessment model by using a case study of an integrated green infrastructure site. In the Sydney region the Parramatta CBD area has been chosen to test the proposed model.

The overall research interest lies in examining how green infrastructure is currently being assessed for sustainability, revealing the gaps in current research and subsequent application in practice, and proposing and testing potential indicator-based model.

Following this approach, the overarching research question is: What are the criteria for designing a comprehensive and integrated assessment model that measures the sustainability performance of green infrastructure networks and is able to do so in ways that are both comprehensive (addressing economic, social, and ecological factors) and integrated (dealing with the nature of green infrastructures as interconnected networks with multiple purposes)?

Based on the overall research question above, and the research objectives, eight specific research questions related to these objectives have been identified. Five research questions relate to the first specific objective:

1. What is the research philosophy (conceptual framework)? 2. What are the indicators in use? 3. Are they sufficient to assess and measure the performance of green infrastructure networks comprehensively and in an integrated manner? 4. What are the interactions amongst indicators? 5. What is the order of importance of each indicator? That is, is there a hierarchy amongst indicators in any given assessment ‘system’? Research questions that relate to the second specific objective are: 1. How effective is each indicator in its actual measurement of sustainability? 2. When actually assessing the degree of sustainability of a given green infrastructure, how does it compare to its baseline/benchmark indicator? There is one research question that relates to the third specific objective: 1. What are the strengths and limitations of the proposed assessment model?

1.3 Methodological framework

This research will employ a combination of qualitative and quantitative research methodologies. There are several methods for evaluating the degree of sustainability of urban ecosystems. However, it remains unclear how to bring natural science-based landscape functions together with more comprehensive estimates of all services delivered by ecosystems (de Groot et al. 2010; Nelson et al. 2009; Ranganathan et al. 2008; Volk et al. 2008). The indicator-based model and the composite indices model are the most common models used to assess the current, and to predict future, sustainable urban development. These models have various limitations such as data availability and time/budget required for data collection on the

18 one hand, and spatial and temporal coverage issues on the other. According to Mayer (2008,287):

“… all indices are problematic, if data are unavailable for the majority of the aggregated indicators, which at present is a common weakness to all sustainability efforts regardless of scale or publicity”.

In order to develop a numeric output model, the conceptual framework for the underpinning of this research needs to be defined. The conceptual framework for this research was developed based on the literature and was validated further via semi-structured interviews and an online questionnaire process targeting practitioners, government and academic experts. This framework takes into consideration a combination of ecosystem services, ecosystem health, and human health and well-being. This step is crucial because it determines what is going to be measured, why it is being measured and who will ultimately use the model. Thus, the research aims to establish an indicator-based model.

As defined by Gasparatos (2010,1616), “a composite index is an aggregation of different indicators under a well-developed and pre-determined methodology” (Figure 1.2). An indicator-based approach facilitates many purposes. First, it identifies and analyses the relevant issues, current states and future developments. Secondly, it provides an essential information base for the established objectives, goals, baselines and actions required. Finally, it directs decision makers and planners in terms of monitoring, assessing performance and then controlling/regulating. This approach can also be employed as an effective method to communicate amongst stakeholders (designers and the public communities) to raise their awareness about sustainability status of their surrounding environment (Dizdaroglu 2013).

The method utilised to develop an indicators-based model in this study is based on two guidebooks ‘Constructing composite indicators: methodology and user guide proposed’ by (Joint Research Centre-European Commission 2008) and ‘Tools for Composite Indicators Building’ by Nardo et al. (2005). Developing the model in this study involves the following six steps:

1. Defining the problem and developing a theoretical / conceptual framework

A theoretical framework provides a basis for determining the relevant indicators that describe the measured phenomenon. The first step in model development is to identify the purpose of the model. What are the questions being addressed and the objectives for which the model is to be used?

Hence, in accordance with the objective of this research, the DPSIR (Driving force–Pressure– State–Impact–Response) model was utilised. According to Niemeijer and de Groot (2008), the DPSIR framework is the most common method for developing sustainability indicators

19 worldwide. It helps to clarify the complex relationship between cause and effect variables, to understand the issues behind variability, and to identify potential solutions. The proposed DPSIR framework is addressed in Chapter 2 following the literature review.

2. Selecting indicators and screening

This step involves the selection of the indicators that are linked to the theoretical/conceptual framework. An indicator is a quantitative measure of relevant phenomena that envisions current conditions, or changes, in order to set goals, strategies and solutions (Heink and Kowarik 2010). This step is seeking to respond to the following questions: What is to be included within the scope of the model? What are the important variables and parameters? What drives the system? What outputs will be generated? This is addressed more fully in Chapter 4 and 5. These indicators show both quantitative and qualitative information that will simplify the relationship between the key indicators and green infrastructure performance.

For some indicators, information will be specifically quantitative and others may be more qualitative. For example, in estimating the reduction in CO2 emission, energy savings and the air and water quality can be numerically quantified. However, for other indicators, especially those under the social and cultural benefits categories, such as ‘community cohesion’ or ‘health’ or ‘quality of life’, it is more difficult to define parameters and to measure them quantitatively. In these cases, proxy variables will be considered.

3. Imputation of missing data

There are two general methods for dealing with missing data. The first method is case deletion, which is based on omitting the missing data from the analysis. The other method is based on providing a value for missing data. In this method, the missing data values are generated through single imputation (e.g., mean/median/mode substitution), regression imputation, expectation-maximisation imputation, or multiple imputations (e.g., Markov Chain, Monte Carlo algorithm). In this study, the second method (regression imputation) is utilised to deal with missing data.

4. Normalisation

Indicators should be normalised to a comparable unit before weighting and aggregation (Figure 1.2). The commonly used normalisation methods are: (1) ranking, which allows the performance of indicators to be followed over time in terms of relative positions, (2) standardisation, which converts indicators to a common scale with a mean of zero and standard deviation of one, (3) Min-Max, which allows indicators to have an identical range by subtracting the minimum value and dividing by the range of the indicator values; and (4) categorical scale, which assigns a score to each indicator.

In this study, benchmarking values (baselines) ranging between 1 and 5 have been employed. They represent maximum and minimum values and refer to different contribution levels of sustainable green infrastructure for each indicator. These baseline values have been assigned based on the results of other studies as well as some, which are site specific and applicable in Australia. In some cases, the baseline value was proposed specifically based on the study area and can be adjusted in accordance with the project objectives and desired targets set by stakeholders and decision-makers.

Figure 1. 2 Construction of an index (Boulanger 2008, 47)

5. Weighting and aggregation

The weighting procedure establishes the importance given to the indicators by stakeholders. Weighting methods include statistical models (i.e., factor analysis, data envelopment analysis, unobserved components models), and participatory methods (i.e., budget allocation, analytic hierarchy processes). Furthermore, weights can be determined based on expert opinion. Different aggregation methods are possible: summing up (linear aggregation); multiplying (geometric aggregation); or aggregation using non-linear techniques (multi-criteria analysis).

In this study, the budget allocation method was used for category weighting and the Weighted Average Index method (WAI) was utilised for weighting individual indicators in relation to others.

6. Visualisation of the results

This step involves the interpretation of the findings in order to provide a clear and accurate presentation of index results. Many visualisation techniques exist, such as tabular format, bar or line charts, ranking or dashboards. This study represents outputs in a GIS tool. The study area will be divided into 100 x 100 metre-grid cells and all indicator values will be applied to the grid cells. This means that each cell will have its own assessment value. The visualisation process will produce a set of composite index values that will:

- Present green infrastructure performance in each cell and each category. - Provide the possibility to redesign and reassess; and - Determine the sustainability level of the green infrastructure in the entire study area.

In addition, the outputs of this step, that are structured based on policy, strategies and decision support, will offer the potential to consider both present and future sustainability levels. For example, the output will establish an agenda for decision makers to help in:

- Defining various scenarios regarding the planning of green infrastructure capacity for the future. - Establishing long-term sustainability outcomes; and - Identifying the best planning option for future developments.

Therefore, the model reviews the capacity and sustainability level of current green infrastructure performance and allows for predicting future scenarios. However, because of the time constraint, in this study the model is tested only via application to the existing green infrastructure in the study area.

1.4 Research contributions

This study will provide a quantitative assessment pilot model to establish the score value (low to high) of, and the performance of, existing or proposed green infrastructure sites. It will determine whether the site is providing positive or negative results in relation to the established sustainability benchmark values that are explained in Chapter 6. The model will generate results that can provide tangible information to city-planners, urban developers, architects, civil engineers and landscape architects, enabling them to understand the value of green infrastructure and compare its sustainable performance to that of conventional infrastructure. This will help industry professionals make decisions that will shape the well-being of cities and their inhabitants and determine the value of the benefits of sustainable urban development. This project will make a considerable contribution to the fields of landscape, urban planning and land development.

This research has constructed the first model to measure green infrastructure performance by integrating three evaluation frameworks: ecosystem services, ecosystem health, and human health. This model may contribute to a framework for green infrastructure rating and specifically as a sustainable GI rating system tool in Australia. Recently, AILA (Australia Institute of Landscape Architecture) has had the intention to adopt and develop the US rating tools for landscape design, SITES, in Australia - the development of a rating tool for the assessment of landscape infrastructure. This research also has the potential to be of significant value as a policy tool in helping to achieve measured sustainability targets and outcomes.

1.5 Thesis outline

This thesis is structured into eight chapters and three appendixes: an overview and introduction, literature review, methodology, results from the semi-structured interviews and online questionnaire, results from spatial analysis, and, finally, a discussion of the results, followed by conclusions.

 Chapter 1- Introduction This chapter describes the background of the research, research gaps, research objectives and questions.

 Chapter 2- Literature Review: Key Concepts of green infrastructure This chapter presents a review of the relevant literature beginning with the scope, concept, and definitions of green infrastructure, and discussion of the literature on sustainable urban development as well as ecosystem services. This is followed by a review of existing methods, models and tools that assess the performance of landscape components in various aspects in order to determine the gaps in existing studies. The chapter concludes with a proposed conceptual framework for assessing green infrastructure performance based on literature review.

 Chapter 3 - Methodology

This chapter provides a detailed description of the research methodology and discusses how the research was developed and conducted. A description of the methodology used to address the study aims and objectives is also provided. The mixed methods approach applied in this research is explained.

 Chapter 4- Semi-structured interviews analysis

Chapter 4 explains the qualitative research method (semi-structured interview) that is used to extract expert attitudes on green infrastructure definitions, types, structure, concepts and framework as well as to verify the conceptual framework which is proposed, based on the literature review presented in Chapter 2. This chapter also identifies initial list of indicators

23 based on literature review and experts’ knowledge and experiences derived from the semi- structured interviews.

 Chapter 5 - Online questionnaire analysis

To develop a composite index model, it was required to select key indicators from the initial list of indicators. To accomplish this task, a quantitative research method (online questionnaire) is used to rate the importance or “weight” of indicators as described in Chapter 5. Next, the indicators are normalised, weighted and aggregated based on the OECD (2008) guidebook.

 Chapter 6- Develop assessment metrics  Chapter 7- Application of case study

In Chapter 6 equations and baseline values for all indicator are described. The model is calibrated through a pilot study of the Parramatta CBD, which is a major urban area in the geographic heart of the Sydney metropolitan region (Chapter 7). This study area has good potential in testing the model because it includes mixed land use with a variety of green infrastructure types, configurations and distributions as well as data availability. In addition, Parramatta City Council is currently developing detailed green-space planning for this part of the city called the ‘Green Grid Master Plan’, that will incorporate green infrastructure as part of its design strategy. In the testing and building of the model, a number of software programs are employed such as ENVI, ArcGIS and iTree Eco. Visualisation results are illustrated in Chapter 7.

 Chapter 8: Discussion and Conclusion

This chapter provides a discussion of the sustainability performance of the study area with reference to the proposed model, as well as a presentation of its limitations and weaknesses. It concludes by way of a discussion of whether the findings of this study achieve the objectives and answer all the research questions. Additionally, it suggests areas for further research and development.

 Appendix A: Questionnaire form  Appendix B: Questionnaire results analysis  Appendix C: Frequency of species and calculation of Shannon diversity score

Publications As part of the validation of the methodology that was undertaken in the study, findings from each method have been presented at several conferences and published in conference proceedings and scientific journals. Results from semi-structured interviews were presented at the State of Australian Cities (SOAC) conference in 2015 and published in the SOAC conference proceedings:

1) Pakzad, P. & Osmond, P. 2015a. A conceptual framework for assessing green infrastructure sustainability performance in Australia. State of Australian Cities (SOAC) 2015, Gold Coast, Queensland.

Development of initial list of indicators and conceptual framework were presented at the Urban Planning and Architectural Design for Sustainable Development (UPADSD) conference in 2015 in Italy, and were published in the Social and Behavioral Science Procedia in 2016:

2) Pakzad, P. & Osmond, P. 2016. Developing a sustainability indicator set for measuring green infrastructure performance. Procedia - Social and Behavioral Sciences, 216, 68-79.

A result from online questionnaire was presented at the International High Performance Built Environments conference in 2016 and will be published in the Engineering Procedia in 2017: 3) Pakzad, P. & Osmond, P. & Corkery, L. 2017. Developing key sustainability indicators for assessing green infrastructure performance. International High Performance Built Environments conference, 17-18 November 2016, Sydney. Procedia Engineering, 180, 146-156.

GIS modelling (Carbon and Urban Heat Island)

4) Pakzad, P. & Osmond, P. & Philipp, C. H. 2015. Review of tools for quantifying the contribution of green infrastructure to carbon performance. ICUC9 - 9th International Conference on Urban Climate jointly with 12th Symposium on the Urban Environment, 20- 24 July 2015, France.

5) Philipp, C. H., Wannous, J. & Pakzad, P. 2015. Thermal impact of blue infrastructure: Case study Cheonggyecheon, Seoul (Korea). ICUC9 - 9th International Conference on Urban Climate jointly with 12th Symposium on the Urban Environment, 20-24 July 2015, France.

CHAPTER 2: LITERATURE REVIEW

Introduction This chapter presents the review of the literature and relevant studies and has seven sections. The first section describes the concept and terminology of green infrastructure. The second and third sections explain the history of GI from the past to the present followed by a GI typology and the key principles of GI planning. The fourth section reviews the existing tools, models and guidelines for measuring the sustainability performance of green infrastructure. This section aims to critically assess their weaknesses and strengths and highlights significant gaps in their assessment methodologies. The fifth section describes the existing GI frameworks followed by describing the widely applied DPSIR framework and its associated parameters (section 6). The chapter concludes with the proposed integrated conceptual framework based on the supporting theories and existing models.

Green Infrastructure (GI)

Green infrastructure is a broad concept (Table 2.1) that is applied and used in various disciplines (Davies et al. 2006). It can be a system for water management (Ahern 2010); it can refer to ecological services and land conservation; or it can be an attitude for a strategic planning approach (McDonald et al. 2005). Green infrastructure is an emerging concept based on ecological systems and their ecosystem services. A variety of perceptions exist among scholars, institutions and governments about definitions, implications and implementations. They extend from “contributing to land conservation and providing clean water to enhancing territorial cohesion” (Dige 2011 ,27). The following sections describe a number of existing definitions and terminologies, brief history, characteristics and components of green infrastructure types before deriving a conceptual framework.

2.1 Green infrastructure concept and terminology

Over the last two decades the term “green infrastructure” has become a popular term in city and regional planning theories and practices where it is focused towards sustainable land use (Ahern 2007; Beatley 2000, 2012; Eisenman 2013; Enlow 2002; Gill et al. 2007; Hansen and Pauleit 2014; Mazza et al. 2011; Mell 2016; Sanesi et al. 2017). It has become one of the ingredients for formulating sustainable urban development policy and management. The concept of green infrastructure as a “life support system” originated in the United States in the 1990s. A number of cities in America, for example, in the states of Pennsylvania, Florida and Maryland, have incorporated green infrastructure concepts into their city planning policies (Benedict and McMahon 2006). They have increasingly taken on the challenge of managing storm water runoff, water conservation and water pollution reduction through the implementation and use of green infrastructure (Flynn 2011).

Whilst there is no single definition for green infrastructure, it is generally recognised as “… all the natural, semi-natural and artificial networks of multifunctional ecological systems within, around and between urban areas at all spatial scales” (Tzoulas et al. 2007 ,6). However, another definition outlines green infrastructure as a network in nature that supports ecological functions and ecosystem services. Benedict and McMahon (2006) presented a similar view on green infrastructure and defined it as an interconnected network of natural, semi-natural and engineered components consummate with each other to conserve and support an ecosystem service and human well-being. Table (2.1) presents number of definitions of green infrastructure.

Table 2. 1 Definitions of green infrastructure in various disciplines

Definitions Key aspect Reference

GI is the physical environment within and between cities, towns and villages. The network of open spaces, waterways, gardens, woodlands, Social, economic and green corridors, street trees and open countryside that brings many (TEP 2005) environmental benefits social, economic and environmental benefits to local people and communities.

An interconnected network of natural areas and other open spaces that conserve the natural ecosystem values and functions, sustains clean air Conservation (Benedict and and water, and provides a wide array of benefits to people and wildlife. McMahon 2006)

Green infrastructure is the network of natural places and systems in, around and beyond urban areas. It includes trees, parks, gardens, Recreation allotments, cemeteries, woodlands, green corridors, rivers and (CABE 2011)

wetlands provide recreational green space.

GI is a network of natural landscape assets which underpin the (Australian Institute economic, socio-cultural and environmental functionality of our cities Multifunctional and life of Landscape and towns – i.e. the green spaces and water systems which intersperse, support for human and other Architect (AILA) connect, and provide vital life support for humans and other species species 2012) within our urban environments.

Green infrastructure maintains and improves ecological functions in combination with multifunctional land uses. Natural and human-made Ecological functions (European structures or a territory devoid of permanent human-made structures Commission 2009) that provide — directly or indirectly, partly or totally through the vegetation it supports, a series of services to society.

Connections between natural sites and habitats. Valuable green urban Species migration and (European areas and human-made bridges to natural areas, ecological corridors habitat connectivity Environment and zones where habitats merge. Agency 2011)

Green infrastructure is a strategic approach to land conservation, a 'smart' conservation that addresses the ecological and social impacts of Conservation (Benedict and sprawl and the accelerated consumption and fragmentation of open McMahon 2002) land.

Green infrastructure is an approach to stormwater management that (The US uses soils and vegetation to utilise, enhance and/or mimic the natural Stormwater management Environmental hydrologic cycle processes of infiltration, evapotranspiration and Protection Agency reuse. (EPA) 2016)

GI is a concept that is principally structured by a hybrid hydrological/drainage network, complementing and linking hydraulic Linkage of hydrological green areas with built infrastructure that provides ecological functions. network and providing (Ahern 2007) It is the principles of landscape ecology applied to urban environments. ecological functions

GI is a strategically planned and delivered network of high-quality green spaces and other environmental features. It should be designed (NECR126 Natural and managed as a multifunctional resource capable of delivering a wide Improving quality of life England range of environmental benefits for local communities and improving Commissioned quality of life. Green infrastructure includes parks, open spaces, 2013) playing fields, woodlands, allotments and private gardens.

The interconnection, and subsequent integration, of individual components into networks is a key principle of green infrastructure (Benedict and McMahon 2006; Kambites and Owen 2006; Pauleit et al. 2011). This networking combines different ecosystem services, leading to what some have called the multi-functionality of green infrastructure (Coutts 2016; Lafortezza et al. 2013; MacFarlane 2007; Pauleit et al. 2011; Tiwary et al. 2016). Hence, green infrastructure can be recognised as a natural life-support system that provides a variety of benefits to people, ecosystems and the economy (Pitman et al. 2015). Simultaneously, green infrastructure preserves land, conserves habitat, restores the natural environment and minimises habitat fragmentation. In this way, multi-functionality represents an important contribution in the reconceptualisation of green infrastructure, from a single system into an integrated network (Figure 2.1).

Figure 2. 1 A schematic green network in a city

The European Commision (2012) classifies green infrastructure functions in four prime groups including: (1) protecting ecosystem state and biodiversity; (2) improving ecosystem

28 functioning and promoting ecosystem services; (3) promoting societal well-being and health; and (4) supporting the development of a green economy and sustainable land and water management. Table (2.2) summarises key benefits of green infrastructure that are defined by scholars.

Table 2. 2 Key benefits of green infrastructure in literature

Sources Key benefits  Sustain clean air and water Benedict and McMahon (2006)  Conserve natural ecosystem

 Maintenance of biodiversity Breuste et al. (2008)

Carter and Butler  Protecting functional landscape (2008)  Providing many social, economic and environmental benefits to local people and communities  Supporting biodiversity and native species Mell (2008); Young  Sustaining air and water resources (2011)  Contributing to the health and enhancing quality of life  Maintaining natural ecological processes

 Providing socio-economic benefits  Providing healthy environments Yli-Pelkonen (2008)  Improving physical and mental health

 Accessibility and walkability of urban GI that improve physical, psychological and Abraham et al. (2010) social health and well-being

 Providing walkable neighbourhoods  Providing multiple economic and social benefits Wolch et al. (2010)  Ecological benefits: Biodiversity conservation and runoff infiltration

 Cooling effects of urban area Carter (2011)  Flood control

 Economic benefits of GI: e.g. avoid cost by not-commuting by car, health effects Vandermeulen et al. from cycling, the value of lower emissions, improved traffic safety (2011)

 Runoff control Wong et al. (2012)  Regulation (Water flow regulation, Natural hazard and disaster regulation, Climate regulation)  Cultural (Aesthetic values, Recreation and religious use, Job and employment Schäffler and Swilling creation, Education and public awareness raising) (2013)  Provisioning (e.g. Food provision,)  Supporting services (e.g. Soil formation, Photosynthesis, Nutrient cycling, Biodiversity and habitats protection)

 Adaptation to climate change  Delivering ecosystem services Douglas (2012)  Providing health benefits for human

 Improving human health and well-being  Providing a wide range of natural functions and ecosystem services (cleaner air and Pitman and Ely (2013) water and healthier soils)  Contribute to climate change adaptation and mitigation strategies

 Conserving biodiversity and urban habitat  Food productions  Economic benefits

 Enhancing biodiversity  Providing food productions  Microclimate control Lovell and Taylor (2013)  Carbon sequestration  Improving visual quality  Improving physical activity and social capital

Di Leo et al. (2016);  Moderating outdoor air temperature (urban heat island mitigation) and providing Norton et al. (2015); human thermal comfort Wang et al. (2015)  Improving individual and public health Coutts et al. (2012);  Conserve natural ecosystem services and functions Coutts (2016)  Sustain clean air and water

2.2 History of green infrastructure thinking

The concept of green infrastructure has emerged from a long line of thinking. There are historical precedents across several fields of study that provide the basis for the recent theories that have been developed. Green infrastructure is rooted in the past: it stems from the ideas of parkways, the British garden cities, greenbelts, greenways, ecological corridors and smart growth (Youngquist 2009). Benedict and McMahon (2006) declared that the term green infrastructure is relatively new, but the concept is not: the so-called “old wine in new bottles” (MacFarlane 2007 ,156). Table 2.3 illustrates the rise of green infrastructure movement.

Table 2. 3 Green infrastructure movement milestones

Time period Milestones

1850-1900 Scenic and historic approach: Relationship between humans and nature

1900-1950s Land and resource conservation approach (land and nature ethic): Greenway, greenbelt and garden city movement, considering land conservation and preservation value

1960-1970s Ecological approach: Greenways for considering ecological value (ecological connectivity)

1970-1980s Recreational, social and cultural approach: Greenways for considering value of connecting people and place (social cohesion and recreation value)

1990-present Comprehensive approach: Sustainable urban development, smart growth and major infrastructure projects that incorporate green infrastructure

The foundations of green infrastructure thinking go back to the mid- to late- 19th Century (Figure 2.2). In this period, the relationship between humans and nature changed dramatically in the context of the industrial revolution and rapid urban growth that resulted in different settlement patterns that incorporated what today we know as green infrastructure. This includes Olmsted’s planned green suburb of Riverside near Chicago, Illinois (Benedict and McMahon 2006; Olmsted et al. 1868), and Howard’s Garden City of To-Morrow (Mell 2010).

Figure 2. 2 History of green infrastructure planning

Central Park in New York was designed by Olmsted in 1857 and is considered as one of the first green infrastructure projects. This project was followed by others including Prospect Park (1859) in Brooklyn and Franklin Park in Boston in 1878. The Emerald Necklace in Boston was the first stream corridor (blue infrastructure) in the USA - also designed by Olmsted. There are further works by Charles Elliot that connected Franklin Park and a number of other public spaces at the city scale in Boston and Detroit. These networks have delivered multifunctional facilities, recreation opportunities and flood control (Figure 2.3). Network and interconnectivity were part of the philosophical thinking of Olmsted that parks should be linked to one another and surrounding neighbourhoods (Ahern 1995) to provide “peace of mind” and “enlarged sense of freedom” while preserving the nature and wildlife habitat (Benedict and McMahon 2006 ,26) and they were highlighted in his projects. As Davies and others have stated, green infrastructure may be viewed as ‘Olmstedian values’ of innovation and connectivity (Davies et al. 2006 ,43; Mell 2010).

Figure 2. 3 Emerald Necklace map, Franklin Park, Boston, Massachusetts

In the United Kingdom, in 1898, Sir Ebenezer Howard promoted the Garden City movement. Garden cities were intended to be planned, self-contained communities surrounded by greenbelts, containing proportionate areas of residences, industry and agriculture (Figure 2.4). Howard’s theories for comprehensive city planning and green belt design are reflected in the development of the two towns, Letchworth (1904) and Welwyn Garden City (1920), along the rail network connecting to London. These towns were surrounded by parklands, farmlands and greenbelt to limit city growth and to provide recreational areas for residents. Nowadays, Howard's egalitarian ideas of the Garden City model are influential across many nations, including Australia (Canberra). In 1928, the British Garden City model was implemented in the USA. Examples are represented in Radburn, the Woodbourne neighbourhood of Boston, the towns Greenbelt in Maryland, Greenhills in Ohio, and Greendale in Wisconsin. In all these cases, on both sides of the Atlantic, these patterns were designed by professional practitioners.

Figure 2. 4 Ebenezer Howard's vision of a garden city and social city (1898)

In the early twentieth century, there was significant concern for land conservation and resource preservation amongst various disciplines, appearing in what have become seminal texts. This gave rise to land ethics that became part of the national agenda in the United States. The term ‘land ethic’ was coined by Aldo Leopold. It is a philosophy that considers the actions when humans use or make changes to the land. Aldo Leopold is best known for his book A Sand County Almanac (1949). He was influential in the development of modern environmental ‘land ethics’, ‘ethical dealing’ and ‘the movement for wilderness conservation’ (Leopold 1949). This was several decades before ecological principles became applied to urban planning and landscape design. In this era, natural environments not only provided recreation and enjoyment for people, they also were acknowledged for providing ethics for biodiversity and species conservation and protection.

In 1939, a German geographer, Carl Troll, coined the term ‘landscape ecology’ to describe a new field of knowledge for applying the interpretation of aerial photography to studies of the distribution of landscape elements, the flow of energy and environment (Wu and Hobbs 2007). Ecological urbanism is another fundamental concept of green infrastructure in urban design and planning. It is an integrative approach of landscape and ecology to the design and management of cities and built environment. In 1969, McHarg, in his book, Design with Nature, introduced the idea of ‘physiographic determinism’ and a new meaningful method using overlays to evaluate this approach. He emphasised the importance of natural features in designing and encouraged planners and decision makers to consider an environmentally conscious approach to land design (McHarg 1969).

Another important thinker on the ecological approach to city planning and urban nature is McHarg’s student, and later colleague, Anne Whiston Spirn. Her book, written in 1985, ‘The Granite Garden: Urban Nature and Human Design’ argued that cities are part of the natural world and demonstrated how cities can be designed accordingly based on natural processes (Spirn 1985). In Canada, Hough took the same approach as McHarg and Spirn and published his book in 1984 entitled ‘City Form and Natural Process’.

About 1960, the landscape architect Philip Lewis was among those who began to focus on the importance of understanding the potential of land and began designing ecological, recreational and cultural corridors in his projects. His study, and practical experiments, focused on the preservation of ecological corridors to deliver environmental, public health and biodiversity. Kevin Lynch, in his last book, “Wasting Away” in 1990, undertook an ecological approach to managing resources and waste. He stressed the roles that natural features played in increasing identity, coherence, legibility and immediacy of urban form.

More recently, in the book ‘Greenways for America’ (1990), Charles Little stated that the term greenway originated from a cross between the ‘parkway’ from nineteenth century in the USA and the British term ‘greenbelt’. Little emphasised that a greenway is defined as linear corridors for mobility and access to be used by pedestrians and as bicycle routes across the city (Austin 2014). Also, he presented greenways in five categories: “Urban riparian corridors, recreational

33 greenways, ecological corridors, scenic and historic routes and comprehensive networks” (Mell 2010 ,43). Benedict and McMahon (2006) and Austin (2014) use the same categorisation.

By the 1990s, sustainable development was becoming a national and international goal. In the USA, new ways of looking at the landscape have evolved as a result of the term “green infrastructure” becoming widespread in practice, theory and in literature. Numerous authors have written extensively on the benefits that green infrastructure may deliver. At that time, green infrastructure was defined as a linear landscape (greenway) in undeveloped land which is shared concurrently by people and wildlife. In 1999, the President’s Council on Sustainable Development (USA) officially defined a new vision of sustainable future development (Spitzer 1999, 141). The report outlined that green infrastructure was an essential part of the community and should “conserve, protect, restore and manage the local and regional networks of natural living environmental resources and amenities” (Benedict and McMahon 2006 ,35).

Harris and Noss at the University of Florida were amongst the preeminent scientists to advocate the concept of ecological and trail networks, landscape connectivity and integrated conservation systems. They brought the green infrastructure concept into action through the ecological and recreational networks in Florida’s master planning (Noss and Cooperrider 1994; Noss et al. 1997). Noss proposed that the regional reserve system comprised core preserves, corridors and buffer zones (inner and outer) (Figure 2.5).

Figure 2. 5 The regional reserve system proposed by Larry Harris and Reed Noss (Noss and Cooperrider 1994)

The Florida network was designed in a GIS-based model including both an ecological network and a social/cultural network. This model has been used to identify potential hubs and linkages, to conserve native landscapes and to maintain and restore connectivity amongst native ecosystems and processes while also connecting urban area to networks of trails that enhance human benefits (Figure 2.6).

Figure 2. 6 Florida GIS-based model: Greenways System Planning Project, 1999 (Benedict and McMahon 2006, 36)

Davies et al. (2006) and CABE (2011) presented green infrastructure in structural terms as a physical environment that lies within and extends beyond our cities, towns and villages. The characteristics of multi-functionality (Sandstrom 2002; Tzoulas et al. 2007), interconnectivity and the integrity of habitat systems and green parcels of land (Van der Ryn and Cowan 1996) that encompass all environmental resources contribute towards sustainable resource management. That is achieved by providing a framework for economic growth and nature conservation .

Brady et al. (2001) defined green infrastructure as natural resources and features of urban and suburban landscapes such as trees, streams, wetlands, and open spaces. Weber et al. (2006), Tzoulas et al. (2007) and Lafortezza et al. (2013) outlined green infrastructure as a strategic way to combine and analyse linkages between ecosystem services and human health and well- being. In a broad formulation, according to Benedict and McMahon (2006), green infrastructure is the ecological framework needed for environmental, social, and economic sustainability.

Green infrastructure approach can also improve resiliency in the face of climate change overtime. ‘Resilience’ refers to the capacity of a system to absorb changes while still retaining the same function and structure (Holling 1973; Walker et al. 2004) and this leads to what has been termed in literature as ‘ecological resilience’. In this sense, adaptive capacity of green infrastructure components support the processes that modify ecological resilience. Dige (2011), Coutts (2016) and Derkzen et al. (2017) extend this argument by highlighting how green infrastructure capacities contribute to the adaptation to, and mitigation of, climate change and urban resiliency through the protection and provision of ecosystem services. From the

35 perspective of climate change and its effects on the environment, the term green infrastructure has appeared increasingly in land management, planning and land capacity. In England, the National Planning Policy of 2012 (Department of Communities and Local Government, UK, 2012) drew up a new framework for climate change adaptation and mitigation through the planning of green infrastructure in the city (Lafortezza et al. 2013).

The United Nations in 2005 published a holistic report called the Millennium Ecosystem Assessment. It was an important international collaborative effort to map and conserve the Earth’s ecosystems. This report applies to all scales of green infrastructure from the global to the local. It provides a general framework for classifying the benefits of Ecosystem Services into four different groups which can structure the green infrastructure performance framework (Ely and Pitman 2012 ,35). These categories are: (1) Provisioning services, products obtained from ecosystems that provide food, fibre, fuel and materials; (2) Cultural services, nonmaterial benefits obtained from ecosystems that provide aesthetic and psychological benefits, recreational activities and related health benefits, and a sense of place (3) Regulating services, benefits obtained from regulation of ecosystem processes that moderate environmental conditions and quality such as stormwater regulation, pollution clean-up, carbon dioxide sequestration, energy and local climate change control; and (4) Supporting services that underline all ecosystem services such as soil formation, nutrient cycling and primary production.

The European Commission (2009,16) identifies green infrastructure as a strategic planning approach that maintains ecological functions at the landscape scale in combination with multifunctional land uses. This EU definition reflects an emerging consensus amongst scholars and international agencies that the notion of an ‘ecological framework’ is a key concept of green infrastructure. An ecosystem is a dynamic and complex system of ‘plant, animal and microorganism communities and nonliving environment interacting as a functional unit’ (Millennium Ecosystem Assessment 2005a, 49). Natural cycles (Water, Carbon, Nitrogen, and Mineral cycles) are attributed to ecosystem services that deliver clean air and fresh water, generate food, regenerate soil, regulate climate and sequester carbon (Weber et al. 2006). These are known as the essential nutrients for a healthy life and a sustainable environment.

To conclude, the concept of green infrastructure has evolved in parallel with the evolution of the professions and disciplines of landscape architecture, landscape planning and urban and regional planning over the decades from the middle of the 19th century. It is underpinned by the concept of ecosystem services (TEP 2005). This contributes to the conservation of biological diversity (Tzoulas et al. 2007). The provision of ecosystem services is a fundamental concept for understanding the performance of green infrastructure. Ecosystem services provide a variety of benefits through the transformation of resources (environmental assets) such as land, water, soil, vegetation and the atmosphere into goods and services like clean water, air, food and energy (Ely and Pitman 2012). Many of these ecosystem services are provided through green infrastructure. Therefore the ‘ecological framework’ is an umbrella, which embodies all ecosystem services to sustain ecosystem health and human health and well-being.

Ecosystem services and their parameters, which are related to performance of green infrastructure, will be explored in section 6 of this Chapter.

2.3 Green infrastructure components and key principles

A holistic view of green infrastructure incorporates both the natural environment and engineered systems to provide clean water, conserve ecosystem values and functions, and provide a wide array of benefits to people and wildlife. It can be applied to different scales. On the local level, green infrastructure practices include, but are not limited to, rain gardens, permeable pavements, green roofs, infiltration planters, trees and tree boxes, and rainwater harvesting systems. At the largest scale, the preservation and restoration of natural landscapes (such as forests, floodplains and wetlands) are critical components of green infrastructure. For example, Ely and Pitman classified green infrastructure components into public parks, gardens (public and private), green roofs and walls, squares and plazas, greenways including river and creek corridors and green routes along major transport corridors (road, rail, and tram) (Ely and Pitman 2012 ,29).

On the other hand, the NECR126 (Natural England Commissioned Reports, 2013, 1) has outlined green infrastructure assets differently, according to common categories of vegetation: (1) Street trees (2) Hedges (3) Grassland (4) Woodland (5) Ponds (6) Grass verges (7) Gardens and parks including cemeteries (8) Green walls (9) Green roofs (10) Rivers and canals and (11) Urban stormwater drainage systems. Moreover, green infrastructure types are not all focused on vegetation – it includes rivers and streams, often also categorised as ‘blue infrastructure’.

According to Benedict and McMahon (2006,13 & 127), the three main physical components of green infrastructure networks are hubs, corridors (links) and sites. These three components illustrate the pattern of the green infrastructure. They vary in function, shape, size and ownership (Figure 2.7). Some are public conservation lands; others are private lands.

Hubs are large blocks of the least-fragmented ecological landscape that ‘anchor’ the network. They provide space for native plants and animal communities and an origin or destination for wildlife, people, water, nutrients, and energy moving through the system. Hubs need to be large enough to work as suitable ecological building blocks for the network. Sites are smaller than hubs and may not be attached to the network, but like the two other components can provide important ecological and social values to the green infrastructure network as a whole. Corridors connect these hubs, provide linkage and tie the system together, providing conduits for movement for plants, animals, water, and nutrients. Corridors also mitigate the effects of habitat fragmentation and enhance the overall resiliency of natural systems in relation to natural and human disturbances. It is more effective and efficient to connect hubs with appropriate land uses on a scientific basis rather than connecting nodes between two hubs randomly.

Figure 2. 7 Green infrastructure components (Hubs, Corridors and sites)

Ian Mell has consolidated GI principles into seven general categories based on a common- sense analysis of green infrastructure potential benefits identified by other scholars (Mell 2008 ,5). According to Mell, the intellectual root of green infrastructure:

“draws its foundation from the ideas of Greenways (connectivity, access), Smart Growth (integrated development) and Urban Greening (sustainable design). The concept of green infrastructure has utilised the focus and drive of these disciplines to develop a broader planning remit” (Mell 2008, 3).

Table 2.4 shows Mell’s principles, but here they are reduced to four categories. Three categories – Integration across boundaries, Ecological resource conservation and Multiple benefits – overlapped with other categories and have been integrated into the Connectivity and Multi-functionality groups (modified by author).

Table 2. 4 Green infrastructure key principles (Mell 2008, 73; revised by author) Key principles Description

Integrated and connected networks within and beyond GI planning aims for added values derived from interlinking the components green spaces functionally and physically while integrating with other types of urban infrastructure and built environment.

Multi-functionality and multiple benefits GI provides several ecological, socio-cultural, and economic benefits concurrently. It means that GI planning aims at intertwining or combining different functions to enhance the capacity of urban greenery to deliver valuable goods and services.

Multiple scales GI planning can be considered for different spatial levels ranging from city-regions to local and neighborhood scale.

Accessible GI planning should provide a good access to activities, resources, services for residents in an appropriate distance. This provides benefits for people by enabling their use; and connecting habitat to enhance biodiversity and limit habitat fragmentation and improve movement between ecological hubs and sites for a variety of species. Source: This table was adapted from Mell (2008, 73). Originally Mell classified seven categories. Here it is reduced to four.

Therefore, the idea of green infrastructure brings together concepts from landscape ecology, human geography, and the planning and design disciplines across different scales (Mell 2008). Green infrastructure affords a comprehensive approach to planning and design to optimise ecosystem services and ecological benefits. Benedict and McMahon (2002) indicate that the ‘greenprint’ approach should follow six principles whose number and intent mirror Mell’s ideas. These principles are: (1) consider multiple functions and benefits to nature and people, (2) consider multiple scales, (3) link hubs into networks to protect biodiversity, (4) use green infrastructure to frame both conservation and development, (5) green infrastructure should be planned and protected before development and (6) incorporate green infrastructure into land- use planning theory and practice based on scientific criteria.

In addition, Benedict and McMahon (2006, 37) stated four more key principles: (7) green infrastructure should be publicly funded in the same way as built infrastructure, (8) green infrastructure respects the needs and desires of various stakeholders in the public, private and nonprofit sectors and reflecting the citizen’s concerns into the design (partnership of land-use system), (9) green infrastucture requires making connections between stakeholders by engaging relevant comunities to activities and programs, and (10) green infrastructure requires long term commitment and thinking that can respond to political changes and relevant legislation and how to ensure a program has the momentum to survive changes in government.

There has been a convergence of thinking on the principles, characteristics and components of green infrastructure in less than two decades. This convergence is summarised in the synthesis diagram in Figure 2.8.

Figure 2. 8 Synthesis diagram of green infrastructure

2.4 Review of existing GI assessment models and applications

This section reviews and analyses existing methods for assessing green infrastructure with the intent of providing a coherent and integrated framework for the continued development of a more refined assessment tool.

The role of green infrastructure is to provide a variety of benefits to society and to nature. These benefits are underpinned by their ecosystem services. These benefits are increasingly recognised as important criteria in urban planning and design decision-making agendas. It follows that decision makers must be able to select the appropriate assessment method to properly evaluate green infrastructure options.

Currently, there are several tools that can be applied to evaluate the benefits and services of green infrastructure. Thus, it is important to identify and assess which tools are the most appropriate to use in defining and calculating the benefits inherent in green infrastructure. Over the past decade, a number of tools have been developed for measuring and rating the sustainability performance of buildings and technical infrastructure. Only a few tools have been specifically developed to measure the performance of green infrastructure. By considering all

40 the benefits that green infrastructure provides in the ecological, social and economic arenas, it is beneficial that an empirical tool is specifically designed to assess its effectiveness, costs, and benefits accordingly.

The existing tools are designed to accomplish various objectives and purposes. For example, the ‘Health Economic Assessment Tool (HEAT)’ for walking and cycling estimates the benefits of doing exercise regularly in terms of reducing mortality risk. The ‘Green Infrastructure Valuation Toolkit (GIVT)’ converts the number of ecosystem services provided by green infrastructure to a monetary value. The Heliwell system addresses the visual amenity (aesthetics) of trees (NECR126 Natural England Commissioned 2013). Therefore, it is crucial to know precisely about the tools’ objectives and their contexts to enable users to achieve a valid assessment.

To make this comparative assessment, two steps have been undertaken. Firstly, 16 infrastructure assessment tools have been identified in the literature review and are listed in Table 2.5. This compiles the current known tools for assessing the benefits of green infrastructure. The table addresses three criteria: (1) what is the tool’s origin? (2) What does the tool assess? (3) In which contexts can it be applied? Secondly, the characteristics of these tools are placed into a matrix. The characteristics have been synthesized from the analysis of the literature. One axis contains the three types of nature services (provisioning, cultural and regulating services) and biodiversity. The other axis lists the green infrastructure elements (types). Then tools from Table (2.5) allocate to cells in Table (2.6) (in intersection of axes), based on their coverage and attributes. This analysis is then presented in Table 2.6. It shows the lack of a single, comprehensive tool that can perform a comprehensive and integrated evaluation of a site, single green infrastructure network, or a network of networks. It also reveals which services and features are not supported by a given assessment tool.

Table 2. 5 Green infrastructure evaluation tools

Name, sponsor and Applicability Characteristics References date of Origin Guide to Valuing Green roofs, - This is a free guide which provides an annual (Center for Green Trees, Rain monetary and quantitative value of four benefits of Neighbourhoo Infrastructure gardens, GI across all spatial scales. d Technology Permeable - Measures economic value by using valuation (CNT) 2010) Center for pavements, methods based on scientific research Neighbourhood Water - The tool defines GI as a ‘network of decentralised Technology (CNT) harvesting stormwater management practices’ (NECR 2013, 49) Chicago, USA, - Uses ecosystem services as a basis: 2010 (1) Regulation services: climate regulation, water regulation, air quality regulation (decline in NO2, SO2,O3,PM10 due to green roofs, trees and bio-infiltration) (2) Cultural services: recreation and ecotourism. This tool also values energy conservation resulting from green roofs and trees

Name, sponsor and Applicability Characteristics References date of Origin - The tool calculates annual benefits, so discounting rates need to be applied. Integrated Terrestrial and - INVEST is a free series of 16 computer models. It is (NECR126 Valuation of Riverine open-source software to map, quantify and value Natural Environmental systems such as ecosystem goods and services under different England Services and Parks, scenarios of land, water and marine uses. It can be Commissione Tradeoffs grassland, used as part of a stakeholder participation process. d 2013) (InVEST) woodlands - This tool is based on scientific research and has been rivers and tested by experts. It has several publications in peer The Natural Capital wetlands reviewed journals. Project - It provides output in both biophysical and monetary terms. Output helps designers, governments and USA, 2008 decision makers to understand how their decisions will impact on environments and human well-being - It is not designed for a specific location and can be applied and adapted to anywhere with available data. - It is a multi-disciplinary modelling tool which applies at all scales (local, regional and national). It models marine and terrestrial ecosystem services and multiple alternative scenarios can be compared. - It estimates: (1) Biodiversity (2) Regulatory services: (carbon dioxide sequestration and carbon storage, water purification, sediment retention) (3) Provisioning services: (timber production) (4) supporting services: (crops valued in monetary terms) - It has some limitations including that users should have expertise in ArcGIS. Feedbacks between ecosystem services are not modeled properly and in general, changes in human behaviour are not modeled. GreenSave Green roofs - Web-enabled tool (registration required) easy to use (ATHENA Calculator (GSC) for quick assessment of green roof benefits. Institute 2007) - Estimates the life-cycle cost and compares the cost Green Roofs for of various green roof types and conventional roofs. Healthy Cities, and - The tool is intended to help users compare different the Athena Institute alternatives over a time period (1 to 60 years) to (for US and determine which one has lower cost in lifespan. It Canada) estimated also future operating, maintenance, repair or replacement costs, as well as the benefits of USA, 2009 conditions such as energy savings. - This tool convert benefits of green roofs to dollar values by estimating the energy use reduction in buildings to compare with conventional roofs. Green Tree cover, - Excel spreadsheet tool. (CABE 2011) Infrastructure- Woodland, - Comprehensive cost and benefits of green Valuation Toolkit green roofs and infrastructure assessment toolkit only applicable in (GIVT) urban green UK. However, this tool needs input from economic spaces, green experts through occasional updating of default Commission for corridor values. Architecture and the elements - The benefit estimation tools are grouped into 11 Built Environment (pedestrian and categories which were designed through a series of cycle routes) workshops in the UK. This is a mixed list of UK, 2009 and water ecosystem services and contribution of GI to local economic growth. Include:

Name, sponsor and Applicability Characteristics References date of Origin courses, ponds (1) Climate change adaptation and mitigation, and wetland (2) Labour productivity, (3) Water and flood management, (4) Tourism, (5) Place and communities, (6) Recreation and leisure, (7) Health and well-being, (8) Biodiversity, (9) Land and property values, (10) Land management, (11) Investment - It estimates of the total benefits and full costs to society and, where feasible, they are all expressed in monetary terms to arrive at a net benefit or cost. The strength of this tool is its transparency. Users can understand the process of evaluation. - It estimates the pedestrian and cycle route benefits by estimating the number of people who are encouraged for doing exercise. It will result in reduction of mortality rate. - Weaknesses: significant risk of double counting and overlapping if different benefits such as health, recreation, tourism and labour are aggregated. The tool is a mix of benefits and multiple values. Unit values used in this tool are not proved with evidence and it is not clear how to sum up and or distinguish between different values. - Some group categories like “place and communities” are very large concepts and this tool only addresses a proportion of its indicators. Green Values Street trees, - Web-enabled modelling tool to compare (Center for stormwater green roofs, performance, cost and benefits of GI or Low Impact Neighbourhoo Calculator (GVC) rain gardens, Development (LID), to conventional stormwater d Technology vegetated practices in a single site. It calculates runoff volume (CNT) 2010; Center for swales, trees, reduction (no peak flow). Environmenta Neighbourhood native - For data inputs, landscape details and cost elements l Protection Technology and vegetation and are needed such as size of site, impervious cover, Agency(EPA) (EPA) United States permeable site hydrology and soil details, street and sidewalk 2012) Environmental pavement design and considers life cycle to calculate the Protection Agency volume and peak discharge of runoff produced by that site. This tool is limited to site level. USA- 2004 - GVC can apply up to six green infrastructure Best Management Practices (BMPs) to the scenario, thereby creating a comparison between “conventional” (i.e. pipes, curbs, gutters, and detention ponds) and “green” scenarios. It provides a quantified analysis of LID environmental benefits including reduced runoff volume and maintenance savings, in addition to carbon sequestration, reduced energy use, and groundwater recharge. - GVC displays hydrologic and financial results of the two scenarios side-by-side, highlighting the differences in runoff reductions and financial performance. - It covers Regulating services including carbon dioxide sequestration, storm water reduction and energy use reduction and groundwater recharge in

Name, sponsor and Applicability Characteristics References date of Origin addition to reduced construction and maintenance costs and extended design life. - This tool is unique because it considers construction, annual maintenance costs and full life cycle but it is only applicable in US climatic zones.

Artificial Forest - ARIES is an open-source modelling framework (Bagstad et al. Intelligence for Watershed using artificial intelligence techniques, including 2013) Ecosystem Urban park machine reasoning and pattern recognition, with a Services (ARIES) library of ecosystem service models and spatial data to pair locally appropriate data and models, National Science quantifying ecosystem service flows and their Foundation at the uncertainty within a freely accessible web browser University of and stand-alone software tool. Vermont’s Gund - Like InVEST, scenarios can be modelled, ecosystem Institute for service trade-offs compared, and monetary values Ecological can be applied to biophysical outputs to derive dollar Economics values for some services. USA-2012 (1) Carbon sequestration and storage (2) Flood regulation (3) Coastal flood regulation (4) Aesthetic views and open space proximity (5) Freshwater supply (6) Sediment regulation (7) Subsistence fisheries (8) Recreation The Center for Individual tree - Free excel spreadsheet. (Lal and Urban Forest (Urban trees) - Geographical coverage: U.S. Augustin Research Tree - Estimates CO2 sequestration and building 2011) Carbon Calculator heating/cooling energy saving provided by (CTCC) individual trees. - Climate regulation: Carbon and energy USDA Forest - Output is: Carbon store in the tree, CO2 sequestered Service, Pacific during the past year, Annual energy saving in KWh Southwest Research of electricity and MBtu of heating per tree, also CO2 Station, the Center equivalents of these energy savings. for Urban Forest Research (CUFR). Developed in partnership with the California Department of Forestry and Fire Protection USA-2007 i-Tree Canopy Tree cover - It offers a quick and easy way to produce a (The United statistically valid estimate of land cover types (e.g., States USDA (United tree cover) using aerial images available in Google Department of States Department Maps. Agriculture of Agriculture) - The data can be used by urban forest managers to 2006) Forest Service estimate tree canopy cover, set canopy goals, and track success; and to estimate inputs for use in i- USA-2006 Tree Hydro and elsewhere where land cover data are needed.

Name, sponsor and Applicability Characteristics References date of Origin i-Tree Streets - It focuses on the ecosystem services provided and structure of street tree populations. US-2012 - It is used by the community to estimate economic values of tree performance to manage annual budget and determines which trees provide more benefits of importance to the community. It encourages publicity to support street trees and to increase the quality of life in neighbourhood to all urban scales. - It makes use of a sample or complete inventory to quantify and put a dollar value on the trees’ annual environmental and aesthetic benefits, including energy conservation, air quality improvement, carbon dioxide reduction, stormwater control, and property value increases. - It covers (1) Regulating services including carbon dioxide reduction, storm water reduction and (2) Cultural services: Aesthetics as captured by property value increase, local climate control and energy consumption due to tree shade.

i-Tree Design - Evaluates the ecosystem services of a tree to parcel level. - Link to Google Maps for the USA and Canada, USA and Canada- by selecting a tree near a property it shows the 2012 effect of the selected tree on the energy use and represents the other benefits. - It covers ecosystem service including:

(1) Regulatory services; climate regulation: carbon dioxide, (2) Water regulation: storm water capture, (3) Air quality regulation: air pollution, (4) Energy conservation due to tree shade.

i-Tree Hydro - It is the first vegetation-specific urban (Beta) hydrology model. - It is designed to model the effects of changes in USA-2012 urban tree cover and impervious surfaces on hourly stream flows and water quality at the watershed level. i-Tree Eco - It estimates ecological benefits of trees from (ENSPEC single tree to any size project. 2012; The USA-2010 - It covers ecosystem services including; United States Australia 2012 - (1) Regulating services: carbon dioxide Department of reduction, storm water capture and air pollution Agriculture and energy conservation due to tree shade and 2006) (2) Cultural services: public health and several

Name, sponsor and Applicability Characteristics References date of Origin biophysical data results that support these (NECR126 services Natural - ‘It is designed to use field data from complete England inventories or randomly located plots Commissione throughout a community along with local d 2013) hourly air pollution and meteorological data to quantify urban forest structure, environmental effects, and value to communities. It helps community to quantify the species, age and size of trees and ecosystem services that trees provide (NECR 2013, 19) - Although this tool was designed for the USA it can be adapted in other countries by changing the input data by ecologists. In 2012 i-Tree Eco was developed in an Australia context by Arboriculture Australia. (ENSPEC 2012) i-Tree Landscape - It allows users to make use of freely available (The United national land cover data maps to assess their States USA-2009 community’s land cover, including tree canopy Department of and some of the ecosystem services provided Agriculture by their current urban forest. 2006) - The effects of planting scenarios on future benefits can also be modelled.

European Health Urban green - HEAT is an online assessment tool to estimate the (World Health Economic spaces, (garden, economic savings resulting from reductions in Organization Assessment Tools parks), mortality as a consequence of regular cycling and/or 2011) (Heat) for walking grasslands, walking. and cycling Pedestrian and - HEAT calculates the answer to the following cycle routs question: if x people cycle or walk y distance on World Health most days, what is the economic value of mortality Organisation rate improvements? (WHO) - It is based on the scientific research and best available evidence, with parameters that can be Europe, 2011 adapted to fit specific situations. Default parameters are valid for the European context. But it can change by users. So this tool is applicable to other countries. - A number of default values provided in HEAT including: mortality rate, value of a statistical life, time period over which you wish average benefits to be calculated and a discount rate, if so wished. - This tool does not cover any specific green infrastructure features. It can apply to any GI that supports opportunities for walking and cycling. MUSIC – model Bioretention - A stormwater modelling tool. It helps urban (eWater 2013) for urban systems, ponds, stormwater professionals visualise and compare stormwater wetlands, design strategies to manage urban stormwater improvement buffer strip hydrology and pollution impacts. It is not a free conceptualisation software. - It is based on scientific research and it can apply in a Monash University wide range of catchments. and Cooperative - It has ability to model additional pollutants; assess Research Centre development submissions; understand peak flow impacts, the effects of storage and detention

Name, sponsor and Applicability Characteristics References date of Origin Australia, 2001 treatments and the water balance; simulate rainwater and stormwater harvesting options and import and export to other models. - The authors have released a version of MUSIC for the United Kingdom. SITES (The - The first sustainability rating system for Landscapes. (SITES 2009) Sustainable Sites - Ecosystem services formed the framework of this InitiativeTM) rating system. - It established a star-based rating system which American Society includes 51 credits organised in nine sections: of Landscape (1) Site Selection Architects (ASLA), (2) Pre-Design Assessment and the Lady Bird Planning Johnson Wildflower (3) Site Design—Water Center at The (4) Site Design—Soil and Vegetation University of Texas (5) Site Design—Materials Selection at Austin and the (6) Site Design—Human Health and United States Well-Being Botanic Garden (7) Construction (USBG) (8) Operations and Maintenance (9) Monitoring and Innovation USA- 2008 Total: 250 points - One Star: 100 points (40% of total points) - Two Stars: 125 points (50% of total points) - Three Stars: 150 points (60% of total points) - Four Stars: 200 points (80% of total points)

The two key outcomes that emerge from this inventory, are that the existing tools and methods for measuring green infrastructure are not addressing all the aspects of ecosystem services and that they do not cover all the features of green infrastructure. Furthermore, the existing tools and the appraisal models for measuring the benefits of green infrastructure are limited by geographic scale, local context and time scale. In some applications, practitioners would need to use more than one tool to undertake a holistic analysis of the design. This is not always practical because the tools utilise different units and valuations. It requires the assessor to establish a common metric or the result may be incorrectly interpreted or inflated.

Among all these tools, three are more inclusive than the others in terms of scale, indicators, and multi-dimensionality. These tools are: i-Tree (for ecological service valuation), the Green Infrastructure Valuation model (for economic cost-benefit assessment), and SITES (a rating system for the level of sustainability of landscapes).

These tools can be adapted to fit the required circumstances if the user has a firm grasp of the principles behind them and the local circumstances to which they are adapted. Otherwise they may have their integrity compromised, and the findings become misleading. An example is i- Tree which was designed for application in the United States and has been used in Australia as well. The following section introduces these three tools in more detail.

2.4.1 Detailed analysis of three green infrastructure evaluation tools i-Tree (Ecological Benefits Assessment Model) i-Tree is an advanced, peer-reviewed software program that was developed by the United States Department of Agriculture’s Forest Service. This tool provides urban and community forestry analysis and benefits assessment. i-Tree STRATUM (Street Tree Resource Analysis Tool for Urban Forest Managers)- recently known as iTree Street is an economic model that is freely accessible. This tool quantifies the environmental benefits of the urban forest and compares this with the management costs. The model can evaluate a range of the benefit factors and ecosystem services that result from trees such as CO2 reduction, air quality improvement, energy conservation, stormwater reduction and property value increases (Ely and Pitman 2012). i-Tree is a successful and state-of-the-art application in the United States that can be used for all scales of development from individual trees to large land parcels, neighbourhoods, cities, and entire states by understanding the local, tangible ecosystem services/benefits that trees provide in the US. The i-Tree model includes six specific analytical tools which provide more detailed information on the contributions of trees to the environment. These are i-Tree Eco, i- Tree Canopy, i-Tree Streets, i-Tree Landscape, i-Tree Design and i-Tree Hydro. (See Table 2.5) Each of its products assesses one part of the ecological benefits of green infrastructure. For example, i-Tree Hydro is focused on hydrology, hydraulic modelling, run-off reduction, and water quality in various land cover types; whereas i-Tree Eco quantifies the urban forest structure and function based on pollution removal, carbon dioxide sequestration, and carbon storage.

The application i-Tree Canopy is a quick method to estimate the overall information about land cover and tree canopy. Data is inputted from a Google map image and the reports are easily outputted to i-Tree Hydro or similar applications where land cover data are needed. i-Tree Streets focuses on the benefits provided by street trees. It converts annual environmental and aesthetic benefits to a dollar value. This tool was developed to help managers and researchers to understand the urban forest structure and designing for the future. It covers ecosystem services such as “Regulating services includes Climate regulation (carbon dioxide reduction), Water regulation (storm water reduction), Cultural services (aesthetics as captured by property value increase in property values), Local climate control, and Energy consumption due to tree shade” (NECR126 Natural England Commissioned 2013, 21). i-Tree Street is now incorporated within iTree Eco and can be adapted to other countries. i-Tree Hydro is a “new application to simulate the effects of changes in tree and the impervious cover characteristics within the watershed stream flow and the water quality” (Ely and Pitman 2012 ,167).

48 i-Tree Landscape (known as i-Tree Vue formerly) is a modelling tool to define the future benefits of urban trees through evaluating the different planting scenarios and assessing the land cover. This includes providing benefits from the current tree canopy and some of the ecosystem services based on the input available from national land cover data maps. i-Tree Design is a simple online tool for the USA and Canada. This tool links to Google maps and by selecting a given property, it illustrates how trees in terms of their size and location, will have an effect on that area. It covers regulation services as a part of ecosystem services including:

1. Climate regulation (carbon dioxide), Water regulation (storm water capture), and Air quality regulation (Air pollution) 2. Energy conservation due to tree shade

The i-Tree Eco program reports on the amount of total carbon stored and the net carbon sequestered annually by trees, the hourly pollution removal rate of the urban forest for ozone, sulphur dioxide, nitrogen dioxide, carbon monoxide, particulate matter (PM2.5) and the annual rainfall captured by trees and the effect of trees on building energy usage and the related reductions in carbon dioxide emissions. i-Tree Eco assesses a complete tree and shrub inventories or random samples of trees on plots for any sized area. These can be analysed and scaled up to address the whole area. This tool measures ecosystem services including:

1. Regulating services including Climate regulation (carbon storage and sequestration), Local climate control, Water regulation (storm water management), Air quality regulation (air pollution) 2. Cultural services (public health) i-Tree has published several useful guidebooks for landscape architects. These books support the framework and methodologies behind i-Trees’ tools. These tools are complete in terms of the ecological benefits of the landscape. However, the ecological benefit that is provided by the landscape is only one part of all the benefits of green infrastructure in an urban context.

Unfortunately, i-Tree is a tool that is focused on the United States policies, strategies and climate conditions and it is not always easily adaptable to other countries. In addition, some of these programs such as i-Tree Streets and i-Tree Design are closed tools and the users are not able to access to the equations to change data defaults and options. Now, i-Tree Eco, Canopy and Hydro can be used internationally.

In 2012, i-Tree Eco was developed for the Australian context by The Melbourne Urban Forest Accord Group. This project was driven by Melbourne and Sydney City Council and Arboriculture Australia. Now, i-Tree Eco Version 6.0 is available for use throughout Australia. To support this tool, the developers have published a guidebook with information about the

49 landscape species in Australia. This guide also includes other criteria for species composition, tree cover, tree density, tree health (crown dieback, tree damage), leaf area, leaf biomass, and information on shrubs and ground cover types.

Green Infrastructure- valuation (Cost and Benefits Toolkit) “Green Infrastructure-valuation” is the only comprehensive economic appraisal toolkit which has been developed recently and covers trees, green roofs and urban green space. But like other methods it has limitations that result in incomplete and approximated results, explained below.

The Green Infrastructure Valuation Toolkit was released in 2010 in the United Kingdom. It is a set of individual spreadsheet-based tools that assess the value of green assets across a wide range of criteria in 11 categories which were derived from various workshops in the UK: (1) Climate change adaptation and mitigation, (2) Labour productivity, (3) Water and flood management, (4) Tourism, (5) Place and communities, (6) Recreation and leisure, (7) Health and well-being, (8) Biodiversity, (9) Land and property values, (10) Land management, and (11) Investment.

All the results are given in financial terms. This tool covers a range of mixed indicators of ecosystem services (climate change adaptation, water and flood management, etc.) and the associated economic benefits of green infrastructure (land and property value, investment and land management). In the second part, the economic value and local economic impacts are aggregated; however, the unit values are not always substantiated based on its literature. This will potentially expose a significant risk of double counting. However, ‘the economic impact is only about the traded economy property values, employment, tourism spending etc. The economic value is about economic welfare which goes beyond the traded economy (NECR126 Natural England Commissioned 2013, 46).

The tool is transparent and users are aware of the valuation process. It can be applicable in other countries if the default data assessed by economists is adjusted based on the local and national regulations. In the United Kingdom, this tool is not recommended for economic valuation without reassessment and input from expert economists and the default values need to be updated to the correct values.

Another limitation of this tool is in the green infrastructure asset’s lifespan benefits evaluation. This toolkit selects various timeframes for each indicator. This timescale is typically associated with economic benefits which affect the gross value added. It assesses the cost-benefits of green infrastructure over the periods of 10, 20-25 and 30-50 years. However, this is dependent on the tree’s condition, for example, and a tree’s life span varies widely.

Sometimes urban trees provide meaningful economic and ecological services over 50+ years (McPherson 2003). It is recognised that aging trees do not only provide ecological services after 50 years; but they also provide benefits in a number of other ways. This includes aesthetic and historical values which are translated to increasing property values and more tourist attraction as well as other indirect benefits.

According to McPherson (2003), costs and benefits of species are varied primarily based on the specific species of tree, size, age, leaf surface area, its maintenance, and its overall condition. However, in this tool, the tree species were not distinguished.

Moreover, green infrastructure is often the network(s) of connected natural spaces that provide greater values than separate and fragmented green spaces. However, “Green Infrastructure- valuation toolkit” does not measure the value of the connectivity and integrated network of nature. In other words, the output of this model is the same for both connected and fragmented natural areas because its input data, the number of trees and land areas that they cover, are similar for both connected and fragmented areas. Size, shape and number of corridors, hubs and patches of green infrastructure are important variables in ecological assessment models because larger size tends to support more species and provide more ecological benefits. “Network models”, “Patch and corridor simulation models” and “Patch dynamic models” are examples of ecological models (Forman 1995).

SITES (Rating System for Sustainable Landscape) Sustainability is related to the ability to control a system (social, economic, and environmental). In 2005, a collaborative comprised of the American Society of Landscape Architects, the Lady Bird Johnson Wildflower Centre, the US Green Building Council, the University of Texas at Austin and the United States Botanic Garden started to develop national guidelines and performance benchmarks for sustainable landscape design, construction and maintenance practices. In 2008, they released the pilot rating tool and applied it to various projects in the US. The second version of this application was released in 2014 (SITES 2009).

It covers a wide range of ecosystem services provided by landscapes, including global climate regulation, local climate regulation, air and water cleansing, water supply and regulation, erosion and sediment control, hazard mitigation, pollination, habitat function, waste decomposition and treatment, human health and well-being, food and renewable non-food products, and cultural benefits.

Ecological services formed the main framework of this tool with 51 sub-indicators in nine major sections, including

(1) Site Selection, (2) Pre-Design Assessment and Planning, (3) Site Design—Water, (4) Site Design—Soil and Vegetation, (5) Site Design—Materials Selection, (6) Site Design—Human Health and Well-Being, (7) Construction, 8- Operations and Maintenance, and (8) Monitoring and Innovation.

It is organized using a star rating scheme on a scale of one to four stars from a potential total of 250 assessment points.

One Star: 100 points (40% of total points) Two Stars: 125 points (50% of total points) Three Stars: 150 points (60% of total points) Four Stars: 200 points (80% of total points)

Significantly, this is the first and at the present the only rating system for sustainability performance of landscapes. It was designed to be applicable in the USA. In 2014, the Australian Institute of Landscape Architects (AILA) commenced work on researching and changing this application to suit Australian conditions, but because of time and budget constraints this project is on hold at the moment.

To summarise Table 2.6 compares tools to determine which application is the best for their purpose. This table helps users to choose an appropriate tool on a given site based on their specific assessment purposes and needs given the availability of input data. However, users need to be aware of the limits of selecting specific methods and be must aware of their incommensurability. The most common pitfall of combining various assessment methods relates to their incommensurability due to double counting of factors and displacement of various units. If the users are not aware of these situations the results will be incorrect.

Moreover, the networked nature of green infrastructure has yet to be accounted for in the evaluation systems and therefore can be seen as a priority area for research and development so an effective assessment tool is available.

Based on the evaluation undertaken, it has been determined that despite the wide variety and the vast differences among the evaluation tools reviewed, there is still limit of capacity for these tools to be adapted to different contexts and different countries, even after adjustment for local circumstances.

Table 2. 6 Intersection between the GI features and ecosystem services and biodiversity(Source: Author).

sustainableurban

Hedges grassand retention Bio and

Gardens,Parks,

canals,streams

Pedestrianand

Rivers

Infiltrationfor

Watercourses

Greenwalls Greenroofs

Streettrees Cycle routs

Grasslands

Woodland

drainage

Wetland

Ponds,

verges GI Feature/ Type Yards

Ecosystem creeks,

services

Food Fresh water ARIES InVEST Fuelwood Provisioning ARIES services Fiber InVEST Biochemicals CNT InVEST Genetic resources InVEST ARIES CNT CNT CNT InVEST InVEST GVC InVEST GVC CNT GIVT CNT i-Tree Eco i-Tree Eco Climate regulation GIVT GIVT i-Tree Eco i-Tree Eco i-Tree Eco ARIES GSC GVC ARIES i-Tree Eco CTCC i-Tree Eco i-Tree Eco GIVT i-Tree Eco i-Tree Eco i-Tree Eco CTCC i-Tree Eco Disease regulation Regulating CNT CNT CNT CNT GIVT InVEST services InVEST GVC ARIES MUSIC GVC GVC Water regulation Eco Eco Eco Eco ARIES Eco GIVT GIVT Eco Eco Eco MUSIC Eco Eco i-Tree Eco GIVT Eco InVEST InVEST Water purification InVEST InVEST

i-Tree Eco i-Tree Eco InVEST i-Tree Eco i-Tree Eco i-Tree Eco InVEST i-Tree Eco i-Tree Eco i-Tree Eco i-Tree Eco i-Tree Eco Pollination i-Tree Eco i-Tree Eco Spiritual and

religious CNT CNT GIVT Recreation and GIVT CNT GIVT GIVT CNT HEAT HEAT CNT HEAT CNT ecotourism HEAT GIVT HEAT HEAT ARIES ARIES ARIES Cultural Services CNT CNT Aesthetic ARIES ARIES ARIES CNT ARIES Inspiration CNT CNT CNT CNT CNT Sense of place Cultural Heritage (Community GIVT value) Biodiversity GIVT GIVT GIVT GIVT GIVT

2.5 Green infrastructure conceptual frameworks in the literature

The conceptual framework is a combination of two viewpoints: theoretical; and practice orientated. A combination of elements from theoretical frameworks and planning process guidance can contribute to the scientific discourse on green infrastructure as well as inform practitioners on planning process. This framework helps to explain the interaction between the natural environment and human activities in practice and in theory.

The exponential growth of the human population and the increasing burden on the capacity of the planet as well as human-driven land alteration, transformation and fragmentation will lead to serious environmental phenomena. Therefore, it is crucial to the deeper understanding of the nature of the human relationship with the planet, to prevent possible future exploitation and destructive behaviour caused by governments, professionals and individuals alike.

There are several different approaches to the environmental ethics and the human-nature relationship. Leopold (1949) developed an early approach to the eco-centric position. Other thinkers such as McHarg (1969) developed a philosophy of ‘Design with Nature’, as did James Lovelock in his ‘Gaia’ hypothesis (1979). An Australian scholar, John Passmore, (1974) in his book ‘Man’s Responsibility for Nature’, emphasised the urgent need for changes in human attitude concerning the environment and land ethics.

Thompson (2000, 16) in the Landscape and Sustainability book chapter entitled ‘The ethics of sustainability’ discussed the various human attitudes towards ‘nature’ and in relation to the environmental ethics. He identifies the following approaches:

1) Anthropocentric position including ego-centric (self-interest) and homo-centric (we are responsible for stewardship of nature for human use and enjoyment)

2) Non-anthropocentric position including Bio-centric and Eco-centric

The following section discusses the value of nature-centred (ecocentrism) and human-centred (anthropocentrism) approaches. It goes on to outline where these two attitudes are in conflict or in agreement with each other by explaining the interaction between the natural environment and human activities. This is defined by identifying the links between the characteristics of green infrastructure and the two approaches in practice and in theory. The overall aim of this section is to establish a conceptual framework based on philosophical thinking in literature.

Environmental ethics is defined as the moral relationship between humans and the natural environment (Leopold 1949). Due to the various ethical perceptions around environmental philosophy, there are many conflicts and disputes. Among them, that between anthropocentrism and ecocentrism is recognised as one of the most common ecological moral dilemmas (Kortenkamp and Moore 2001).

The traditional Western view is that nature is predominantly there to serve the needs of humans, as, for example, mentioned in the Bible. This idea of human-centered or anthropocentrism explicitly dictates that:

“humans are the sole bearers of the intrinsic value of the earth and all other living things are there to sustain humanity’s existence” (MacKinnon 2007, 331).

This human selfishness tends to uncontrollably influence and alter nature’s balance. This is considered to be responsible for the environmental crisis of global warming, ozone depletion, water scarcity and the loss of biological diversity. It would appear that most of the urban design professionals are anthropocentrists in their processes and outcomes (Austin 2014). This is because of economic self-interest, political power, social and cultural prejudices which are all involved in the process of decision making, design and implementation.

In contrast ecocentrism, explains a “nature-centered system of values and extends the inherent worth to all living things regardless of their usefulness to humans” (MacKinnon 2007, 336). It takes the perspective that humans have responsibility for all biological life on earth and that every species has equal rights (Austin 2014). Accordingly, the importance of an ecological footprint of human activities, sustainability assessment methods and urban greenery policies has become an emerging field for discussion between ecocentrism scholars, design professionals and policy makers.

However, the contradiction of advancing ecocentric and anthropocentric attitudes simultaneously explains paradoxical environmental ethical decisions. As a result, there is a gap in the philosophical position of what we have built and its impacts on environmental and public health. In order to evaluate the decision that would equally consider both humans’ rights and nature’s rights, stakeholders must weigh up the possible consequences and determine which one should take priority (MacKinnon 2007).

Green infrastructure is conceptualised as a missing link between humans, nature and the built environment. It is a scientific approach to determine the best use of land to support natural processes, ecosystems and infrastructure to improve quality of life and support settlement’s requirements. GI systems, at various scales, can embody a dual rights-based approach, both anthropocentric and ecocentric in values (Figure 2.9). This is achieved by integrating multiple factors such as pollution mitigation, habitat, biodiversity, the quality of life enhancement, food, energy, recreation and scenic values. Therefore, green infrastructure can be a cost-effective and efficient solution that addresses many of the above issues and problems simultaneously (Austin 2014).

Figure 2. 9 Dual rights-based approach (Source: Author)

What follows is a critical examination of the existing frameworks for urban sustainability indicators and a comparison of the existing green infrastructure conceptual models. This will lead to an outcome that proposes a new framework to facilitate the process of selecting green infrastructure performance indicators to best reflect the comprehensive and integrated function of green infrastructure.

2.5.1 Existing frameworks for assessing urban sustainability

Since the concept of sustainable development first became a major concern, a number of methods, frameworks and tools have been developed to assess the state of, or changes to, urban areas in relation to sustainability performance. The pre-dominant method used to assess sustainability is an indicator-based assessment - applied to many scientific fields from the socio-economic sciences to the environmental sciences. Comprehensive lists of urban sustainability indicators have been developed by international and regional organisations, such as the European Foundation (1998), the European Commission on Science, Research and Development (2000), UN Habitat (2004), the European Commission on Energy Environment and Sustainable Development (2004), the United Nations (2007) and the World Bank (2008). In addition, a number of composite sustainability indices have been developed more recently such as the Environmental Sustainability Index (ESI), the Environmental Performance Index (EPI), the Environmental Vulnerability Index (EVI), the Rio to Johannesburg Dashboard of Sustainability and the Well-being of Nations and National Footprint Accounts (Ecological Footprint and Bio-capacity) (SEDAC, 2007).

The development and selection of urban sustainability indicators is a complex process. The most common framework for selecting indicators is the Causal Network (CN) method. The CN framework is a combination of a series of causal loops and feedback loops, such as the pressure–state–response (PSR) framework and its transformations: the driving force–state– response (DSR); and the driving force–pressure–state–impact–response (DPSIR) (Niemeijer

56 and de Groot 2008). The PSR was proposed by the OECD (1993) and is based on the pressure indicators that explain the problems caused by human activities, state indicators that monitor the physical, chemical and biological quality of environment and response indicators that indicate how society responds to environmental changes and concerns (Segnestam 2003).

The European Environment Agency (EEA) extended the PSR framework to ‘Driving force- Pressure-State-Impact-Response’ (DPSIR) that is now the most internationally recognised framework. The ‘Driving force’ indicators underlie the causes (economic sectors and human activities) through ‘Pressures’ (waste, emissions) to ‘States’ (physical, chemical and biological), and ‘Impact’ indicators that express the level of environmental harm to human health, ecosystem health and functionality. Ultimately, the setting of indicators, targets and prioritisations are political ‘Responses’ to environmental problems. These causal networks explain the interaction between human activities and natural resources that demonstrate the sustainability level of urban development. Sustainability assessment provides a fundamental approach to the efficient use of natural resources while adapting to human activities and demands, hence provides an essential tool to understand the physical and natural characteristics of urban areas and settlement activities in terms of their potentials, weaknesses and risks in the urban planning process (Lein 2008).

Integrating the green infrastructure concept into the planning process has important influences to mitigate negative impacts of future development. Figure 2.10 demonstrates the DPSIR framework of the linkages between human activities and green infrastructure performance. This flow diagram helps to clarify the complex relationships between cause and effect variables as well as understanding the issues that change the performance of green infrastructure and help to identify potential solutions. For example, connectivity is a key principle of green infrastructure. Any human activities, such as deforestation and land degradation that change the structure of GI, will result in increasing the percentage of impervious surfaces and consequently disturbing functions of ecosystems as well as it will impact on the health of both ecosystems and humans.

Figure 2. 10 DPSIR framework of linkage between human activities and GI performance (Source author)

2.5.2 Existing green infrastructure conceptual models

Numerous social science research models address the environmental effects on human mental and physical health (Table 2.7). The clear consensus is that green open space and biodiversity contribute positively to improving mental and physical health for urban residents. Pickett et al. (2001) proposed an integrated human ecosystem framework for analysing urban systems in relation to their social, biological and physical aspects. The two interconnected parts of this framework are: (1) The human-social system, which includes social institutions and cycles; (2) The resource system, which consists of cultural and socio-economic resources, and ecosystem structure and processes. Grimm et al. (2000) revised Pickett’s human ecosystem framework based on outcomes of land use and land cover changes on the interactions between social and ecological systems. Even though these two models help to explain the concept of green infrastructure, in general, they do not clearly address the relationships between ecosystems and public health (Tzoulas et al. 2007).

The “Arch of health” was developed by the World Health Organisation (WHO) in 1998. This model illustrates the environmental, cultural, socio-economic, working and living conditions, community, lifestyle and hereditary factors of public health. Paton et al. (2005) combined the “Arch of health” model with developmental principles (social, environmental, organisational and personal factors) and systems theory to enhance application within organisations. In 2003, the Millennium Ecosystem Assessment body established a framework for assessing global ecosystem changes and their impacts on human and ecosystem health. This framework links ecosystem services and human well-being through socio-economic factors. Even though this framework is very broad and includes many parameters, it does not ‘explicitly distinguish between the biological, psychological and epidemiological aspects of health’ (Tzoulas et al. 2007, 21).

A comprehensive model developed by Van Kamp et al (2003) synthesised various factors that affect the quality of life including personal, social, cultural, community, natural environment and built environment as well as economic factors. However, the interrelationships between these factors are not clear. Tzoulas et al. (2007) proposed a framework for green infrastructure in urban areas that provides the ground for linking ecological concepts such as ecosystem health to social concepts of individual or community health. On this basis, Lafortezza et al. (2013) described a framework for green infrastructure planning with five interlinked conceptual components: (1) ecosystem services; (2) biodiversity; (3) social and territorial cohesion; (4) sustainable development, and (5) human well-being. In 2010 Abraham et al conducted a scoping study reviewing over 120 studies examining the health-promoting aspects of natural and designed landscapes (Abraham et al. 2010). The authors identified three dimensions of human health linked to green infrastructure: (1) Mental well-being - landscape as a restorative environment; (2) Physical well-being - walkable landscapes and; (3) Social well-being - landscape as a bonding structure. Table 2.7 summarises the most recent frameworks that link ecosystem and human health.

Table 2. 7 Models and theories linking ecosystem and human health aspects (Tzoulas 2007; updated by author)

Author Model/theory Green infrastructure aspect Human health aspect Freeman (1984) Model of Environmental Physical, social and cultural factors Nervous system and Effects on Mental and illness Physical Health Henwood (2002) Psychosocial Stress and Physical poor environment Chronic anxiety, chronic Health Model stress and high blood pressure Pickett et al. (1997, Human Ecosystem Ecosystem structure and processes Socio-ecological systems 2001), Framework and cultural and socio-economic Grimm et al. (2000) resources WHO (1998) Arch of Health Environmental, cultural, socio- Working and living economic factors conditions, community, lifestyle and hereditary factors Paton et al. (2005) Healthy living and Environmental, cultural, socio- Living and working working model economic factors conditions Millennium Links between Provisioning, ecosystem services, Security, basic resources, Assessment ecosystem services and regulating and cultural health, social relationships, (2003) human well-being and freedom of choice Macintyre et al. Framework based on Air, water, food, infectious Health and human needs (2002) basic human needs diseases, waste disposal, pollution (biological, personal, social, and spiritual) van Kamp et al. Domains of liveability Natural environment, natural Health all aspects (2003) and and quality of life resources, landscapes, flora and (physical, psychological, Circerchia (1996) fauna, green areas social) TEP (2008) Life support system and High-quality natural environment Movement network(Active sustainable growth (environmental capacity), travel mode and impacts Managing surface waters; on human health and well- biodiversity; climate change being); productivity adaptation (Sustaining jobs )

Tzoulas et al. (2007) Conceptual framework Ecosystem services and functions Socio-economic, and Austin (2014) integrating Green (air and water purification, climate community, physical and Infrastructure, and radiation regulation, etc.) and psychological health ecosystem and human ecosystem health (air quality, soil health. structure etc.) Abraham et al. Human health and well- Accessibility, walkability, Physical, psychological (2010) being benefits of green Aesthetically appealing rural green, and social health and well- infrastructure being Environmental aspects (air quality and noise reduction), Biophilia, restorative, social and cultural interactions

2.6 Redefining the GI conceptual framework The DPSIR framework (Figure 2.10) provides the basis to establish a composite indicator- based model for assessing green infrastructure performance. Frequently, in the green infrastructure literature, the concept of ecosystem services is adopted to replace, and explain, the functions and benefits of green infrastructure from the global to the local scales (Tzoulas 2007; Mazza et al. 2011; Lovell et al. 2013; Austin 2014; Hansen et al. 2014; Ely and Pitman 2014). The combination of both green infrastructure and ecosystem services theories into a unified framework seems promising.

Figure 2.11 demonstrates the links between human health and well-being, ecosystem health and ecosystem services. This model respects both philosophical anthropocentrism and ecocentrism. The link between these three systems is very clear. A healthy ecosystem within a green infrastructure environment has the ability to increase the delivery of ecological and cultural services to improve human health and well-being at both individual and community scales. This conceptual framework proposed in Figure 2.11 helps to identify relevant indicators for assessing the performance of green infrastructure.

Figure 2. 11 Conceptual framework of green infrastructure performance assessment

In the following section, each of these three systems is discussed in more detail to explore the link between GI with each of these categories and to set up the initial list of indicators based on this integrated concept for developing the assessment model: ecosystem services parameters, human health parameters and ecosystem health parameters.

2.6.1 Ecosystem services parameters Ecosystem services are defined as “the benefits people obtain from ecosystems” (MEA 2005a; 2005b). Many studies were conducted in this area with different approaches such as Millennium Ecosystem Assessment (MEA, 2005a). This assessment focuses on nature, ecosystem functions and variety of processes and fluxes whilst the TEEB study by Sukhdev et al. (2010) focuses on the economics of ecosystems and biodiversity. There is a consensus in the literature that the concept of ecosystem services underlines not only provisioning services (mainly marketable goods such as food, fodder, wood and raw materials), but also the regulating and cultural services (Table 2.8). This framework indicates the interaction between the natural systems that contribute to human health and well-being and emphasises the economic and non-economic values of ecosystem services and natural capital while self- maintaining the natural system (supporting services) (de Groot et al. 2010; MEA 2005a; TEEB 2010).

Table 2. 8 Millennium Ecosystem Services Framework (Millennium Ecosystem Assessment 2003, 57) Provisioning Services Regulating Services Cultural Services Products obtained from Benefits obtained from Nonmaterial benefits obtained ecosystems regulation of ecosystem from ecosystems processes  Food  Climate regulation  Spiritual and religious  Fresh water  Disease regulation  Recreation and ecotourism  Fuel wood  Water regulation  Aesthetic  Fibre  Water purification  Inspirational  Biochemical  pollination  Educational  Genetic resources  Sense of place  Cultural heritage Supporting Services Services necessary for the production of all other ecosystem services  Soil formation  Nutrient cycling  Primary production

Ecosystem services cover several relevant aspects for multifunctional planning such as how to improve the delivery of services through ecosystems while offering benefits for humans. For example, the framework addresses how to improve the human health, social cohesion and also secure the health of ecological systems and natural processes (Tzoulas et al. 2007; e.g., Chan et al. 2012; Haase et al. 2012; European Commission 2012; Lafortezza et al. 2013; Hansen & Pauleit 2014).

Ecosystem functions are the physical, chemical and biological processes that are necessary for its self-maintenance, which result from the interaction between the biotic and abiotic components of an ecosystem (de Groot et al. 2002; Turner and Chapin 2005). However, Hansen and Pauleit (2014) indicate that functions and services are not substitutable, and “function” is not typically used in the literature refer to ecosystem services, although some scholars use

61 functions in a fuzzy way, often meaning the same as services (E.g. Mazza et al. 2011; Lovell and Taylor 2013).

Consequently, to avoid further confusion that may lead to uncertainty regarding terminology and before developing the green infrastructure conceptual framework for this study, it is crucial to justify this research’s approach for using the terms “ecosystem services, green infrastructure performance, human benefits and natural processes”. To do this the following questions need to be answered:

- Are these terms interchangeable “Functions ≃ Performances ≃ Processes ≃ Services ≃ Benefits”? - Can “ecosystem services” be conceived as “ecosystem benefits” and consequently “green infrastructure benefits”? - How are these terms related to “green infrastructure performance”?

Functions ~ Processes and Services ~ Performances ~ Benefits in the Development of an Assessment Framework

According to Forman and Godron (1986), function is the interaction among the spatial elements and the natural process and flows of energy, materials, and species through the components of an ecosystem. Pacala and Kinzing (2002) take a similar approach and grouped ecosystem functions in three categories as below:

1- Stocks of energy and materials (for example, biomass, genes); 2- Fluxes of energy or material processing (for example, productivity, decomposition); 3- Stability of rates or stocks over time (for example, resilience, predictability).

In addition, the term “capacity” can also be found in the literature about functions, goods and services. Capacity covers both performance and capability, or potential, of an ecosystem or a landscape to deliver services. In the TEEB study by Sukhdev et al. (2010) functions are also considered as purely ecological phenomena and it is the capacity of natural processes and components to provide goods and services that fulfill human needs directly or indirectly (de Groot et al. 2010; de Groot et al. 2002). Indirect functions are known as regulating or internal functions that are important for the existence of ecosystems, regardless of perception by humans (e.g. nutrient cycles, water cycle, and litter decomposition). Direct or social/cultural services satisfy human needs and demands. The Millennium Ecosystem Services Assessment (MEA) framework uses the word “services” for human and ecosystem benefits instead of functions. Jax (2003) suggested using the term “function” exclusively when it is related to humans and delivers a service for human welfare and health. In contrast, Fisher et al. (2009) conveyed that service as per definition only requires human beneficiaries while functions of ecosystems, such as soil formation, may be vital for their existence but not necessarily directly utilised by humans.

Based on the literature review, ecosystem functions are considered a subcategory of ecological processes within ecosystem structures. In turn, natural processes are the result of the complex interaction between biotic (living organisms) and abiotic (chemical and physical) elements of an ecosystem across the driving forces of matter and energy. Figure 2.12 is based on the Ecosystem Services Cascade model proposed by Haines-Young and Potschin (2010). It clearly illustrates the sequence and distinction between functions, services and processes.

Biophysical Structures

Ecosystem Ecosystem Benefits Values Functions Services

Natural Processes

Ecosystems & Biodiversity Human health & Well-being

Figure 2. 12 Cascade model for linking ecosystems to human well-being (Haines-Young and Potschin 2010 and de Groot et al. 2010, 11) de Groot et al. (2002) classified ecosystem functions in four groups: regulation; habitat; information; and production. The regulation function is defined as the maintenance of life support systems and essential processes. It refers to regulating services in the millennium ecosystem services (MEA). The habitat function is defined by the provision of living space and maintenance of biological and genetic diversity (biodiversity). The production function includes the provision of biomass, food and raw materials. It can be referred to provisioning services from MEA. While the information function includes services such as spiritual enrichment, mental development and leisure refers to social and cultural services. The urban forest ESG established an indicator-based framework which was founded on four categories of ecosystem functions proposed by de Groot et al. (2002). Additionally, ecosystem disservices such as damage to infrastructure and building foundations by tree roots were included as an indicator in the ESG framework.

Ely and Pitman (2014) tabulate the ecosystem services that can be provided by green infrastructure (refer to Table 2.9). Their framework is based on the “triple bottom line” of sustainable development (Ely and Pitman 2014, 17) which represents the benefits of green infrastructure in three categories - environmental, social and economic. Their approach includes the more anthropocentric functions of natural environment.

Table 2. 9 Ecosystem services that can be provided by green infrastructure (Ely and Pitman 2012; 2014, 28)

Theme Categories Sub-categories Environmental. Climatic Temperature Shading modification reduction evapotranspiration Wind speed modification Climate change Carbon sequestration and storage mitigation Avoided emissions (reduced energy use) Air quality Pollutant removal improvement Avoided emissions Water cycle Flow control and Canopy interception modification flood reduction Soil infiltration and storage Water quality improvement Soil improvements Soil stabilization Increased permeability Waste decomposition and nutrient cycling. Biodiversity Species diversity Habitat and corridors Food production Productive agricultural land Urban agriculture Social Human health and Physical well-being. Social and psychological Community Cultural Visual and aesthetic Economic Commercial vitality Increased property values Value of ecosystem services However, Austin (2014) explained the contribution of green infrastructure to ecosystem services by demonstrating the interlinkages between ecosystem health and human health and well-being through the framework proposed by Tzoulas et al. (2007). This framework considers both ecocentric and anthropocentric approaches (Table 2.10) and has been developed by the author through adding the natural processes (energy, carbon, water etc.) as a supporting function.

Table 2. 10 GI contributes to ecosystem and human health through services delivered by ecosystem (Noss and Cooperrider 1994; Tzoulas et al. 2007 and Austin 2014; revised by author).

Among these definitions and approaches, this study defines the term “services” in a general way that provides benefits to humans and nature (directly and indirectly). To avoid mixing the functions of green infrastructure and the concept of ecosystem services in the following sections, the term ‘‘services’’ is used while ‘‘functions’’ refer to the ecological functioning of green infrastructure elements.

The synthesis of the two classifications is presented in Tables 2.8 and 2.9. This study defines that services refers to four functions which are covered by green infrastructure:

- Production, - Regulatory, - Cultural/social functions, and - Biodiversity/habitat functions.

Some of these functions are for the benefit of humans but some are for securing the health of the ecosystem, self-maintenance and existence of the system that is not directly utilised by humans. All these functions result from physical, chemical and biological processes through interaction between biotic and abiotic components. This helps to optimise the performance of an entire system. This approach is very similar to what has been presented in Table 2.10.

Performance is defined as the capacity of or the capability of a system. It shows the status of a system or comparison between capability of various systems at the time a decision is taken or during a life span of the system. Therefore, targets for each system and relevant indicators have to be set up. Performance also shows the relationship between transitions from the functions of natural processes to the ideal level or benchmark. The performance of each indicator can range from low to high and it can determine the degree of functionality of a system in relation to its targets and the quality of the intended services that it can or is expected to achieve. In Chapter 4, the performance of green infrastructure and the relevant indicators are investigated and explored in depth.

2.6.2 Human health parameters There is a growing realisation that nature plays a significant role in the creation of sustainable and healthy cities (Irvine et al. 2010). The link between human health, human contact and interaction with nature and its healing power is well-established across a variety of disciplines including biology, phycology, sociology and urbanism (Macintyre et al. 1993; Bartley et al. 1997; Diez et al. 1999, 1997; Dunn and Hayes 2000; Ross 2000; Ulrich 2002; de Vries et al. 2003; Mayer and McPherson 2004; Kim & Kaplan 2004; Irvine et al. 2010). Accordingly, all these studies across multidisciplinary areas have been made to improve the quality of life in neighbourhoods and cities (Diener et al. 2006) through increasing access to natural settings.

Human life encompasses four aspects of existence: physical, mental, spiritual and emotional (Figure 2.13). The physical aspect is related to the status of the physical body and it refers to

65 our ability to survive and thrive in the material world. The mental aspect embraces intelligence and the capacity to think. It covers our thoughts, attitudes, beliefs and values. The spiritual aspect refers to our inner essence, our soul, surrounding by time and space and connects our feeling to the universe. The emotional aspect is the ability to experience life deeply, to relate to and meaningfully connect to one another and also surrounding environment.

Aspects of healthy life Human health

Wholness process Emotional Spritual Mental Phycical

Connect to the Experience Feeling to one Feeling to Religious universe and Experience our Thoughts surronding Healing process another ourselves beliefs spritual body material world essence

Figure 2. 13 Aspects of healthy life (Source: Author, unpublished master thesis)

Human health parameters and green infrastructure

There are a large number of theories that outline the effects of contact with nature on human health and well-being in order to find balance in all four aspects of human health. Abraham Maslow (1970) was a pioneer in defining humans needs across different levels of importance. Based on Maslow’s theory, Macintyre et al (2002) suggested a conceptual framework to show the importance of environment or place to human health. This framework includes various environmental, social and economic factors which influence humans’ health. However, it did not recognise the importance of a habitat’s biodiversity in contributing to the positive outcomes from these factors (Tzoulas et al 2007). Townsend and Ebden (2006) adapted this theory to develop a new framework, named “Feel Blue, Touch Green” in which people who suffer from depression and/or anxiety can achieve the highest level of Maslow’s hierarchy (a sense of spirituality and personal fulfilment) when given the opportunity to interact with nature. Likewise, Swan (1977) and Nebbe (2006) expressed a similar idea in that “the spiritual or enlightening experiences which Maslow (1970) refers to as peak experience are often realised to occur in natural environments” (cited in Townsend et al. 2010, 15).

In the 1980s, the “Biophilia hypothesis” or love of nature and living things was developed by Wilson. It asserts that humans are instinctively drawn to nature. To support this theory, Kellert and Derr (1998) outlined the contributions and values that resulted from contact with nature across the various aspects of physical, emotional and intellectual growth (Townsend et al. 2010, 11):

 Aesthetic value (physical attraction and beauty of nature): adaptability, heightened awareness, harmony, balance, curiosity, exploration, creativity and an antidote to the pressures of modern living.

 Dominionistic value (mastery and control of nature): coping and mastering adversity, capacity to resolve unexpected problems, leading to self-esteem

 Humanistic value (affection and emotional attachment to nature): fondness and attachment, connection and relationship, trust and kinship, co-operation, sociability and ability to develop allegiances.

 Moralistic value (spiritual and ethical importance of nature): understanding of the relationship between human wholeness and the integrity of the natural world, leading to a sense of harmony and logic.

 Naturalistic value (immersion and direct involvement in nature): immersion in the sense of authenticity of the natural rhythms and systems, leading to mental acuity and physical fitness.

 Negativistic value (fear and aversion of nature): developing a healthy respect for the risks, power and dangers inherent in nature with an equivalent sense of awe, reverence and wonder, leading to learning to deal with fears and apprehensions in a constructive way.

 Scientific value (knowledge and understanding of nature): developing a cognitive capacity for critical thinking, analytical abilities, and problem-solving skills leading to competence.

 Symbolic value (metaphorical and figurative significance of nature): being able to access the limitless opportunities offered by the process in the natural world to develop understanding of one’s own circumstances, leading to cognitive growth and adaptability.

 Utilitarian value (material and practical importance of nature): emphasising the practical and material importance of the natural world on which we rely for survival.

Other influential approaches to address the impacts of nature on human health and well-being are “Attention Restoration Theory” (ART) developed by Rachel and Stephen Kaplan (1989) and Ulrich’s “Stress Reduction Theory” (SRT) (Ulrich 1981, 1983; Ulrich et al. 1991). These three theories are closely related to the “Biophilia hypothesis” and “psycho-evolutionary theory”. Attention restoration theory (ART) is mainly concerned with cognitive process and proposes that people can concentrate better after spending time in nature, or even after looking at scenes of nature. Contact with nature can be distinguished on three different levels, where each level has individual benefits for the participant (Pretty 2004; Stone 2006):

 Viewing nature: such as through a window, book, painting or on the television

 Being in the presence of nearby nature: such as walking, cycling to work, reading in the garden or talking to friends in the park

 Active participation in nature: such as gardening, farming, trekking, camping, running, horse riding, hedge laying or forestry

Also, this attention can be voluntary or involuntary according to James (1962, 231). A voluntary attention is a sustained attention that is directed by cognitive-control processes. By contrast, an involuntary attention refers to ‘soft fascination’ that does not require an effort and it is captured by inherently exciting or important stimuli. It is typically associated with contact with nature (Townsend et al 2010) and has the ability to restore attention.

Stress Reduction Theory (SRT) is concerned with the emotional and physiological benefits of exposure to natural spaces and emphasises that natural environments promote recovery from stress and restore expanded energy levels and thereby created an evolutionary advantage to humans (Ulrich 1983; Ulrich et al. 1991). According to Baum et al.’s (1985) definition, Ulrich et al. (1991, 202) regard stress as the ‘‘process by which an individual responds psychologically, physiologically, and often with behaviours, to a situation that challenges or threatens well-being’’.

Referring to Van den Berg et al. (1998), natural elements as “natural tranquilizers” may be particularly beneficial in urban areas where stress is an all too common aspect of daily living. Besides the health benefits that can be gained from green areas, there are also the aesthetic, psychological, ecological, social and economic benefits for society (Chiesura and de Groot 2003). According to Ulrich (2002), being in nature and having connection with it, deeply affects the inner layers of the essence of human beings because natural settings have a tendency to be lower in intensity and less perceptually disordered than many urban environments, they have relatively positive, stress-reducing effects on people (Ulrich & Parsons 1992). Also, only by spending as little as three to five minutes in a natural setting has manifested restorative effects on humans with a combination of physiological, emotional and physical changes (Ulrich 2002).

Pearce and Moran (1994) and later Kettunen et al. (2009) suggest that values that humans obtain from nature can be classified in use-value and non-use value and consequently, direct and indirect benefits. They explain that people value the environment because they use it, both directly and indirectly. Visiting the countryside, walking in the parks or along a riverside are examples of directly using the environment. But natural processes in the environment also provide services to people; these services exist and deliver services to people and other living things constantly. These include climate regulation through absorbing CO2 from the atmosphere, and water quality through natural filtration. Although these are ‘used’ by people, it is indirect as most people are not aware and do not pay attention that they are benefiting in this way.

De Vries (2010) and Ely and Pitman (2014) classified the contribution of green infrastructure to people’s health and well-being in three main categories: 1) physical; 2) psychological and; 3) social health. It demonstrates that physical activities of walking and running, besides

68 providing physical health, deliver psychological and mental health benefits (healing and restoration). Also, it provides more opportunities for casual or structured social interaction through community and group activities (de Vries 2010).

However, Tzoulas et al. (2007) and Austin (2014) have slightly a different classification that includes: 1) socioeconomic; 2) community; 3) physical and; 4) psychological health. They have incorporated the socioeconomic aspect as one of the parameters of human health that are delivered through the implementation of green infrastructure practices.

A study was undertaken by the University of Illinois to investigate the effects of nature on human feelings. This study found that contact with nature reduces mental fatigue and helps to release stress (Kuo et al. 1998). Another study illustrated that small amounts of greenery reduced the levels of inner-city crime and violence and helped to make residential areas a safer environment (Kuo and Sullivan 2001).

Green infrastructure increases recreational opportunities such as walking, running, cycling, sitting and picnicking. A study undertaken in Philadelphia (Stratus 2009) estimated the value of the recreational opportunities increased in “user days” with an increase in the number of green areas. This study is not universal but it shows the relationship between increasing recreational activities and an increased amount of green infrastructure. According to this study, for every “1 additional vegetated acre, there was an increase of 1,340 user days per year and 1 additional vegetated acre provides 27,650 user days over a 40-year period. Therefore, every 1 user day provides $0.71 in present value for 40-year project period. This translates to a benefit of about $951.40 for each additional vegetated acre per year and about $19,631.50 for each additional vegetated acre over a 40-year project period.” (Stratus 2009; cited by CNT 2010).

The network and connectivity of green infrastructure provides networks of formal and informal relationships between neighbourhood residents and supports increased social integration and activities of humans (Coley, Kuo et al. 1997; Kuo, Bacaicoa et al. 1998; Sullivan, et al. 2004). It leads to an improvement in the quality of life and makes a city a better place to live. Also, according to research undertaken by researchers at the University of Illinois at Urbana/Champaign, green cities provide safer and “less violent domestic environments” for residents (Kuo and Sullivan 2001b). Another study reinforced this view and showed that more green spaces reduce the rate of inner-city crime and violence (Kua and Sullivan, 2001a).

In summary, the following is a list of the key social benefits of green infrastructure:

 Improved public health  Improved level of physical activities  Promotes psychological and mental health  Increased community liveability  Improved aesthetic quality

 Increased recreational space  Promotes community cohesion and social interaction  Reduces levels of inner-city crime and violence

2.6.3 Ecosystem health parameters Ecosystem health assessment is now a fundamental factor in ecosystem protection and monitoring. Maintaining a healthy ecosystem is a key management objective for government sectors and environmental agencies. Therefore, a comprehensive and accurate indicator system for ecosystem status is needed (Tang et al. 2013). Due to its complexity, a number of indicators ranging from measurement of single species to composite indicators have been proposed for ecosystem health assessment by ecologists, managers, philosophers, and even economists (Karr et al. 1986; Schaeffer et al. 1988; Rapport 1989; Kay 1991; Haskell et al. 1992; Norton, 1992; Ulanowicz, 1992; Xu et al. 2004 and Beyer et al. 2014). Many attempts have been made to assess the health of various ecosystem types such as lakes, estuaries, lagoons, and agricultural areas (Fonseca et al. 2002; Xu et al. 2005; Xu et al. 2011; Zhang et al. 2007; Zhai et al. 2010 and Brigolin et al. 2014).

Table 2. 11 Studies which have defined ecosystem health (Tzoulas 2007)

Author Type of study Keywords Lu and Li 2003 Model of Vigour index; resilience index; ecosystem health organisation index.

Brussard et al. Discussion of Ecosystem viability or health = 1998 ecosystem current utility, future potential, management containment, resilience.

Lackey 1998 Discussion of Ecological health = ecological ecosystem integrity; need to define the desired management state to achieve desired social benefits.

Costanza 1992 Model of Vigour, organisation, resilience. ecosystem health

The concept of ecosystem health has been defined in a variety of ways (Table 2.11) and the definitions have been closely aligned with the concepts of “stress ecology” (Barrett & Rosenberg 1981; Odum 1985; Rapport 1989) and “ecosystem resilience” theory. Resilience is defined as a capacity in a system to absorb shocks from disturbances that change the system’s ability to return to the state of the same structure and functions (Holling 1973; Walker & Salt 2006). On this basis, an ecosystem can be considered as healthy when it is free from, or resilient to, stress and degradation, and maintains its organisation, productivity and autonomy over time (Costanza 1992; Rapport et al. 1998; Brussard et al. 1998; Karr et al. 1986 (Desha et al. 2016).

Ecosystem health parameters and green infrastructure

Diversity is one of the most important indicators of ecosystem health (Rapport 1995). The link between ecosystem health and public health is the set of ecosystem services provided by the green infrastructure. The benefits of biodiversity for human well-being are generally determined by the diversity of habitats and species in and around urban areas (Tilman 1997). There is a close relationship between ecosystem health and ecosystem services: i.e. increasing ecological stress leads to a reduction in both the quality and quantity of ecological services (Cairns and Pratt 1995). In contrast, healthy ecosystems have the capacity to provide a comprehensive range of ecosystem services (Costanza et al. 1998; Lu and Li 2003). Therefore, ecological functions and ecosystem services derived from green infrastructure can contribute to ecosystem health and to public health, respectively.

The health of an ecosystem is generally defined as the occurrence of normal ecosystem processes and functions (Costanza, 1992). A healthy ecosystem is thought of as one that is free from distress and degradation, maintains its organisation and autonomy over time and is resilient to stress (Costanza 1992; Mageau et al. 1995; Costanza et al. 1998; Rapport et al. 1998; Lu and Li 2003). Some authors have pointed out that defining ecosystem health depends on human-social values and desires (Lackey 1998; Brussard et al. 1998). Therefore, the concept of ecosystem health, like that of human health, integrates numerous ecological, social, economic and political factors.

In summary, the following is a list of the key parameters of ecosystem health:  Air quality  Water quality  Soil quality and structure  Moderate air temperature  Habitat and species diversity  Energy and material cycling

2.7 An integrated approach: Linking GI with ecosystem services, human health and ecosystem health

To conclude, three main approaches have been identified in literature to explain the green infrastructure concept. First, from the ecological and landscape sustainability perspectives, the ecosystem services (or nature services) approach is recognized as a general framework for green infrastructure (Andersson et al. 2014; Austin 2014; Coutts 2016; Lafortezza et al. 2013). It emerges from a global perspective in terms of ecosystem services (ESS), which are delivered by nature, and natural cycles at the global scale which also provide benefits effectively at the local scale (Costanza et al. 2017; Millennium Ecosystem Assessment 2003). Historically, it is linked to the concept of sustainable development and urban ecology (Farr 2011; Hough 2004;

McDonnell and Pickett 1990; Spirn 1985). This ‘services’ approach serves both natural cycles (ecosystem health) as well as human needs (human health) (Alberti et al. 2003; Coutts 2016).

Secondly, from the biodiversity and habitat conservation perspective, linked green spaces or the connectivity approach among networks and systems is another key concept of green infrastructure. It improves ecosystem functionality and it can provide services similar to networks of conventional (hard) engineered infrastructure (Benedict and McMahon 2002).

Thirdly, from an engineering perspective, green infrastructure is considered a specialised form of green engineering, also known as ecological engineering. By substituting conventional infrastructure with living elements, such as vegetation for stormwater runoff reduction and purification, it can deliver and perform ecosystem services and functions.

The main finding from this literature review has been the proposal of a conceptual framework that links green infrastructure, ecosystem and human health and well-being. This framework is seen as a fundamental approach of urban ecology, sustainable landscape planning and urban resilience (Figure 2.14). This framework provides a basis to establish a composite indicator- based model for assessing green infrastructure performance. This is discussed in more details in Chapter 4.

Figure 2. 14 Conceptual framework of green infrastructure based on literature review

Summary

In this chapter, an initial framework based on the literature has been proposed (Figure 2.14) that emphasises the interrelationship between human health at the individual and public levels, ecosystem health and its parameters, and ecosystem services that are delivered by green infrastructure elements across spatial and temporal scales. It is fundamental to the comprehensive integration required to assess green infrastructure holistically.

Following the discussion of methodology in Chapter 3, the proposed framework (Figure 2.14) is validated via a series of interviews with industry, academics and government sector professionals in Australia. The aim of the interviews is to verify the proposed framework and understand representative Australian experts’ attitudes towards GI definitions, concept and types. The outcomes from the interviews conducted with industry representatives are outlined in Chapter 4. After verification of the framework, a set of performance indicators for assessing GI projects has been derived and are also set out in Chapter 4. These indicators then contribute to the development of the composite indicator-based model after weighting and screening via an online questionnaire (Chapter 5).

CHAPTER 3: METHODOLOGY

Introduction

This chapter outlines the research methodology followed to achieve the aim and objectives of the thesis by explaining the methods and sources of data collection, the type of analysis involved and the interpretation of data.

The main purpose of this research has been to develop an indicator-based assessment model as a means to evaluate the performance of urban green infrastructure that is comprehensive, integrated and multi-scale. In order to determine the most effective criteria for such an approach, four major research objectives were formulated:

1. Establishment of a conceptual framework;

2. Investigation of key indicators;

3. Development of a measurement system for indicator-based model; and

4. Validation of the proposed model via a case study.

In accordance with these objectives, a mixed-method approach has been utilised in this study. This chapter highlights the main advantages and disadvantages, and the strengths and weaknesses, of two research methodologies - qualitative and quantitative and two data collection techniques - semi-structured interviews and questionnaire survey. It also provides the rationale for adopting these methods for this study.

3.1 Research design The first research objective was to develop a conceptual framework as a foundation for the model based on a literature review and adjusting it for the Australian context. The literature review has shown that the most accepted framework of GI performance consists of ecosystem services, human health and ecosystem health (Figure 2.13). This framework was utilised as a foundation of this research provided a basis for determining an initial list of indicators to include in the assessment model.

Based on the second research objective, to validate this framework in the Australian context and determine key performance indicators, a mixed-method approach (literature review, interview and questionnaire) was used. An initial list of indicators was proposed by experts through semi-structured interviews and this was reassessed, weighted and screened via an online questionnaire. The results from the semi-structured interviews and online questionnaire are presented in Chapter 4 in more detail. All key performance indicators were formulated

74 based on a review of the technical literature (Chapter 5). Afterwards, spatial analysis was carried out through remote sensing data by applying key indicators on a case study in order to validate the model and visualise the results (Chapter 6). The following section explains the mixed-method approach - qualitative and quantitative and two data collection techniques - semi-structured interviews and questionnaire survey.

3.2 Mixed-method approach

According to Creswell and Garrett (2008), the recent demand for a mixed-method approach arises from concerns about the inability of the individual qualitative and quantitative research methods to cover and address solutions to ever more complex and dynamic problems confronting society. Several definitions exist for mixed methods; one such definition is by Johnson et al. (2007, 123) who defined mixed-methods research as

“the type of research in which a researcher or team of researchers combine elements of qualitative and quantitative research approaches (e.g., viewpoints, data collection, analysis, inference, techniques) for the broad purposes of breadth and depth of understanding and corroboration”.

The diversity in this definition indicates that a mixed-method technique has become critical to good research practice across a variety of project types. According to Greene (2008, 20) the mixed approach “offers deep and potentially inspirational and catalytic opportunities to meaningfully engage with the differences that matter in today’s troubled world”. Underlying these definitions is the recognition of its ability to offer multidimensional research solutions that applying only one form of research method is not able to do. One intention of the mixed- method technique is the diversification of ideas it offers as a concept, coupled with its potential to expand the understanding of human experiences in developing policies and practices.

One major advantage with the mixed method is its strong links to research questions (Creswell and Garrett 2008). According to (Bryman 2015), the decision to choose a mixed-method approach must be based on the purpose of the study, the research questions, and the type of data required for the study. Underlying these reasons is the rationale behind the mixed method in providing the best platform to answer inductive-based and deductive-based research questions together in a single study. Effective utilisation of this principle will yield better outcomes than can be achieved using a single-method approach. For example, combining interviews with a questionnaire survey can help to tap more into participants’ knowledge, yielding powerful insights for the study (Johnson et al. 2007). Equally, the principle behind the mixed method enables the researcher to collect data from multiple sources to investigate the hard and soft issues relating to human and organisational questions without compromising the scientific rigour of the findings (Masadeh 2012). Saunders et al. (2009) and Onwuegbuzie and Johnson (2006) suggest that by adopting qualitative and quantitative research methods within the same research framework, practical questions can be addressed simultaneously from

75 different perspectives, leading to a greater confidence in the findings and conclusions. In addition, adopting a mixed position will enable the researcher to mix and match design elements in a way that provides the best opportunity of answering specific research questions. Table 3.1 presents a detailed account of the mixed-method approach including the two main data collection techniques adopted for this study.

Table 3. 1 Strengths and weaknesses of mixed method approach , adopted from Johnson and Onwuegbuzie, (2004)

Strength Weaknesses Words, images, and description can be used to Can be more expensive to conduct. supplement meaning from figures and vice versa.

Stronger evidence can be provided through convergence Mixing two or more research paradigms can be difficult and corroboration of findings. and problematic.

Can provide broader perspective to a range of research Can be time consuming questions and issues.

Can offer deeper insights and understanding than the Can be difficult to analyse and draw inferences to single approach method. interpret findings.

Can offer more complete knowledge necessary to inform Can generate a large error theory and practice.

3.2.1 Qualitative-method approach

Qualitative-research methods extract relevant information, facts and relevant experiences and knowledge from individuals. This involves the investigation of occurrences from the participants’ viewpoints. As its central objective is linked to the understanding of the participants’ opinions, behaviour, knowledge and experiences, it is regarded as the most suitable way of exploring issues based on social phenomena (Johnson et al. 2007). However, in spite of these numerous benefits, qualitative research has also been criticised for its limited size of data, leading to its inability to provide generalised results (Castro et al. 2010). Such a limited sample size approach has reduced its capacity to illustrate definitive conclusions in obtaining certain research outcomes (Castro et al. 2010; Saunders et al. 2009).

Table 3. 2 Strengths and weaknesses of qualitative and quantitative research methods ((Amaratunga et al. 2002)

Method Strengths Weaknesses

Qualitative  Able to understand people’s meaning  Difficult to control the pace, progress  Able to develop theory. and end-point of research process.  Able to generate data in natural setting  Can be time consuming.  Open data collection approach  Data interpretation can be difficult.  Limited (small) sample

Quantitative  Able to test hypothesis.  Methods used tend to be inflexible  Able to collect large sample. and artificial in nature.  Findings can be generalized.  It is not able to effectively capture human phenomena

Interviews, focus groups and observations are commonly used approaches in qualitative data collection. The most widely used method in qualitative research is the ‘in-depth’ interview where the interviewer can engage the respondent in a dialogue and ask other questions to clarify the respondents’ answers. Various interview techniques are available but the choice of any particular type depends on the nature of the research questions as well as the objectives set out for the study (Gray 2009; Saunders et al. 2009). The most commonly used ones are: structured; semi-structured; and unstructured. Table 3.3 presents the characteristics of the structured, semi- structured and unstructured interviews.

Table 3. 3 The characteristics of interview types (Adapted from Gray 2009)

Structured interview Semi-structured interview Unstructured interview Mainly for quantitative data Mainly for qualitative data Mainly for qualitative data

Capture data speedily Capture data slowly and time Capture data slowly and are time consuming consuming

Uses random sampling Uses purposive sampling Uses purposive sampling

Uses strict interview format Uses flexible interview format or Uses flexible interview format or schedule schedule

Data usually easy to analyse Data may be sometimes difficult to Data usually difficult to analyse analyse

Tend to positivist view of knowledge Mixture of positivist and Mixture of positivist and interpretivist view of knowledge interpretivist view of knowledge

In this study as part of the methodology, semi-structured interviews were conducted with 21 Australian experts who have expertise within the architecture, landscape, urban planning and ecology professions across three sectors of the economy (academic, government and practitioners from industry and consultancy). These interviews were carried out face-to-face and over the phone. The purpose of the interviews was to seek an understanding of the current

77 perceptions and knowledge of GI in the Australian context, and to determine a list of indicators that best evaluate the performance of GI in the built environment.

3.2.2 Quantitative-method approach

A quantitative-research method in the social sciences traditionally involves the measurement of numbers from large amounts of data gathered from various people, often across a large geographical area (Creswell and Garrett 2008). It provides the fundamental connection between an empirical observation and mathematical expression of quantitative relationships (Petty et al.). The measurement of such phenomena, according to Goodman and Santos (2006, 290), helps to “explain the connections between variables by measuring cause-effect relationships”.

A questionnaire can consist of several hierarchical layouts that link to research objectives. The questionnaire may include bot open and closed-ended questions asking participants to choose among various statements that are closest to their own attitude. The Likert scale is a commonly used approach for data collection and was used in this research, by which respondents were asked to indicate agreement on a scale from ‘strongly agree’ to ‘strongly disagree’ (Likert 1967).

Summary This chapter provides an overview of the research methodology that covers a breakdown of the research process, aims and objectives. By combining quantitative and qualitative methodologies, there is a benefit that helps to build a more complete picture of the natural and scientific world (Bryman 2015). Accordingly, with only a limited number of participants (21), a qualitative research method was used to extract views on the understanding of the green infrastructure concept, structure and performance, and to identify performance indicators. In addition, in order to weight the importance of each indicator a quantitative research method was adopted through the subsequent distribution of an online questionnaire to a much broader audience to obtain a larger sample. This is discussed further in Chapter 4. Table 3.3 illustrates the research objectives and the approaches that are used to answer the research questions.

Table 3. 4 Research Objectives, Questions and Approaches

Objective Research question Approach

Literature review

Develop a conceptual framework  What is the research philosophy (conceptual Semi-structured based on the literature review and framework)? interview adjust it in the Australian context

Online questionnaire

Identify and understand the logical  What are the indicators in use? Literature review and rational relationships amongst  Are they sufficient to assess and measure the the various indicators of performance of green infrastructure networks sustainability in relation to green comprehensively and in an integrated manner? Semi-structured infrastructure in order to identify  What are the interactions among indicators? interview the importance of each indicator in  What is the order of importance of each evaluating the performance of indicator? That is, is there a hierarchy among green infrastructure indicators in any given assessment ‘system’? Online questionnaire

Determine and define appropriate Literature review methods that connect and  How effective is each indicator in its actual formulate the indicators which measurement of sustainability? ENVI measure the degree of  When actually assessing the sustainability of a sustainability. This objective will given green infrastructure project, how does it also establish a baseline / compare to its baseline/benchmark? ArcGIS benchmark level for each indicator.  What are the strengths and limitations of the proposed assessment model? iTree Eco

ENVI To test the pilot assessment model by using a case study of an  What are the strengths and limitations of the integrated green infrastructure site ArcGIS in the Sydney region. The proposed assessment model? Parramatta CBD has been chosen to test the proposed model. iTree Eco

CHAPTER 4: INTERVIEW ANALYSIS

Introduction

The proposed conceptual framework based on literature review in the previous Chapter 2 (Figure 2.13) emphasises the interrelationship between human health, at individual and public levels, ecosystem health and its parameters and ecosystem services which are delivered through green infrastructure elements across spatial and temporal scales. The interviews conducted with industry representatives provide further justification for this framework in the Australian context. This section presents the findings from 21 semi-structured interviews with property and design industry professionals (Table 4.1). The dual purposes are: to seek an understanding of the current perceptions and knowledge of green infrastructure in the Australian context; and to refine the draft conceptual framework discussed above to enable the derivation of indicators to evaluate the performance of green infrastructure in the built environment.

The interviewees were selected from the following fields: (1) design (landscape, architecture and urban planning); (2) urban policy; (3) water management; and (4) environment and horticulture - all from the private, academic or government sectors. The interview questions were categorised into three sections: 1. To identify the interviewees’ knowledge of existing tools used for assessing the sustainability performance of buildings, landscape and infrastructure projects. 2. To appreciate their understanding of green infrastructure terminologies, concepts, structure and its components. 3. To identify their priorities for the conceptual framework parameters and interrelations within and between parameters.

The criteria for interviewee selection included their understanding and expertise in sustainability issues, in rating tools and in other assessment methodologies, and their interest and passion for the topic of the performance of green infrastructure.

Table 4. 1 Interviewee profiles

Interview participant type Interviewee numbers Practitioners Building Practitioners 3 Landscape/urban planner/ environmentalist practitioners 10 Academics Landscape and urban planner academics 2 Building academics 1 Government Government official 5 Total 21

4.1 Semi-structured research interview design and data collection approach The most widely employed method in qualitative research is the ‘in-depth’ interview where the interviewer can engage the respondent in a dialogue and ask supplementary questions to clarify the respondents’ answers. Prior to drafting the semi-structured interview questions open question interviews were conducted with four stakeholders. These pilot interviews helped to identify issues that needed to be discussed in more depth with the larger sample. It also provided the opportunity to clarify any ambiguities and to establish more in-depth questions that would direct the respondent to provide further detail, to clarify or to explain answers.

With the knowledge obtained from the literature review and the pilot interviews, a set of 14 questions, including additional probes, were drafted (Table 3.6). Nineteen of the interviews were conducted face to face and two were by phone. The interviews lasted between 45 minutes to one hour and the recordings were supplemented with notes taken during the interview and the impressions, ideas and thoughts of the interviewer. All the interviews were transcribed and manually coded through an Excel spreadsheet. Repetitions, digressions and irrelevant materials were omitted from the transcripts before coding.

4.2 Data analysis

All transcripts were analysed and the frequency of responses for questions that required straightforward “Yes” or “No” answers were counted and tabulated in Table 4.2.

Table 4. 2 Descriptive statistics of answers for Yes/No questions

Sector Government Academics Practitioners Section 1: Establish a general framework based on existing rating tools and assessment methods Q1: Are you familiar with rating tools? Yes (100%) 5 (24%) 3 (14%) 13 (62%) Q2: Have you been involved in developing any rating tools? Yes (38%) 2 (9.5%) 2 (9.5%) 4 (19%) Section 2: Characteristics of rating tools for green infrastructure Q3: Are you familiar with the term green infrastructure? And its definitions? Yes (100%) 5 (24%) 3(14%) 13 (62%) Q4: Do you use any tools or methods for assessing landscape and open space performance as part of a design project? Yes (28.5%) 2 (9.5%) 0 4 (19%) Q4.1 Are you familiar with SITES initiative rating tool? Yes(57%) 1 (5%) 3(14%) 8 (38%) Q6: Do you think that landscape should be assessed as part of an integrated system (green infrastructure) and as an individual component (both)? Yes (62%) 4 (19%) 3 (14%) 6(29%) Section 3: Define framework, categories, sub-categories and benchmarks Q8: Are you familiar with the term ‘Triple-Bottom-Line’ for sustainable development? Yes (100%) 5(24%) 3(14%) 13(62%) Q8.3: Do you think that TBL can be an appropriate concept to create a framework for assessing GI performance? Yes (100%) 5(24%) 3(14%) 13(62%) Q9: Are you familiar with the Millennium Ecosystem Assessment (MEA) framework? Yes (52%) 5(24%) 3(14%) 3(14%)

Q9.1: (After my explanation about MEA); Do you think that MEA can be an appropriate concept to create a framework for assessing GI performance? Yes (71%) 4(19%) 3(14%) 8(38%) Q10: Do you agree with Combining two concepts of MEA and TBL for the measurement of GI performance? Yes (100%) 5(24%) 3(14%) 13(62%) Q11: Do you think a GI assessment framework needs to be benchmarked? Yes (100%) 5(24%) 3(14%) 13(62%) Q12: Do you think benchmarks must cover local government policies and strategies rather than national policies? Yes (19%) 1(5%) 0 3(14%) Note: Only yes and No questions has been considered in this table.

Section 1: Familiarity with existing rating tools

The objective of the questions in this section was to establish whether the interviewee was familiar with any rating tool/s for assessing sustainability in buildings, infrastructure and/or landscape. Eight out of 21 interviewees (38%) were involved in developing rating tools. Six (29%) had provided limited input data at the early stage of the development of a tool and two were involved in the whole development process. The academics had the widest knowledge of the tools available, including the theoretical and philosophical thinking behind them, but they rarely used or tested these tools in practice. These tools are explained briefly in Chapter 2. Table 4.3 demonstrates the order of popularity for the knowledge of the tools.

Table 4. 3 Identifying the most popular tools

IS GreenStar SITES NABERS LEED iTree Others G A P G A P G A P G A P G A P G A P G A P 3 2 4 5 3 10 1 3 8 2 3 6 2 3 9 1 2 6 1 1 5 9 (11%) 18 (23%) 12 (15%) 11 (14%) 14 (18%) 9 (11%) 7 (8%) Note: G: Government; A: Academics: P: Practitioners

Section 2: Understanding of the green infrastructure concept

- GI definitions

All of the participants were familiar with the term ‘green infrastructure’ and they were able to describe it in relation to their own field of expertise. While they were questioned about their understanding of the term, the subjects proposed nine quite different definitions. From these, four broader definitions could be derived. 1: Green infrastructure is a policy and strategic approach to land and species conservation. 2: Green infrastructure is a network of energy, materials and species flows that maintains and improves ecological and environmental functions in combination with multifunctional land uses and provides associated benefits to human populations and ecosystems. 3: GI refers to the combination of ecological functions through the network of natural and engineered systems that are integrated into traditional infrastructure systems to enhance their functions and it can significantly reduce the carbon footprint. 4: GI is an ecological solution underpinned by the concept of ecosystem services to improve the sustainability level of the urban and built environment. It also embraces the social,

82 economic and environmental (Triple Bottom Line of sustainable development-TBL) requirements of the urban environment.

Table 4. 4 GI definitions among three groups of participants

D1 D2 D3 D4 G A P G A P G A P G A P 4 0 2 1 2 0 1 1 5 2 2 8 6 (21%) 3 (11%) 7 (25%) 12 (43%) Note: G: Government; A: Academics: P: Practitioners

Table 4.4 illustrates the frequencies of these green infrastructure definitions amongst the three groups of participants. Some of the participants’ definitions cover more than one category. Most participants defined GI as a solution to improve the sustainability level of the urban environment (definition 4). Interestingly, only three practitioners (11%) were aware of the Millennium Ecosystem Assessment. However, eight of 13 of the practitioners embraced definition 4. Amongst government sector interviewees’ definition 1, the policy and the strategic approach, was favoured over and above other definitions.

- GI components and types

Green infrastructure types vary in terms of scale and functions. Responses from interviewees regarding GI types can be classified into six groups: 1- Green corridors (greenways/ street trees/bike ways) 2- Green roofs and green walls 3- Bio retention and infiltration (bio swales/rain garden/ permeable pavement) 4- Natural green areas (forest/woodland/grassland) 5- Water related components (rivers/ streams/ lakes) 6- Green square/parks/ gardens/ yards

Table 4.5 demonstrates that government sector participants have a broader perspective of GI types at the macro scale than other participants. They were able to name green networks and national parks that have been influential in decision making and strategic planning. Practitioners mostly focused on the micro and meso scales such as green roofs and walls and bio-retention practices. Academics appeared to have a broader perspective across various types of green infrastructure implementation.

Table 4. 5 The most well-known GI’s components

Sector Practitioners Academics Government Green corridors 8 2 5 Green roofs and walls 13 3 1 Bio-retention 8 2 1 Natural green area 3 2 5 Water related components 2 1 3 Parks/Garden/Square 7 3 3

- GI Benefits and functions

Green infrastructure can support numerous individual environmental or socio-economic functions: biodiversity; local distinctiveness; public health, sport and recreation; flood management; and climate change adaptation. They can be 'multifunctional' meaning that different functions or activities occur on the same piece of land (TEP 2008, 9). The interview participants were asked to indicate number of important functions and benefits served by GI. Their responses were classified into 14 categories: 1. property values. 2. visual and aesthetic values 3. recreational opportunities 4. human health and well-being 5. community liveability 6. biodiversity conservation 7. water quality improvement 8. pollutant removal 9. energy efficiency 10. noise reduction 11. carbon sequestration and storage 12. temperature moderation 13. water management (flow control and flood reduction) 14. food production

Figure 4.1 illustrates the relative importance of the potential benefits of GI as designated by the interviewees (normalised between 0-100%).

Figure 4. 1 Importance of potential benefits of GI as designated by stakeholders

Based on Figure 4.1, and the frequencies of participants’ responses to identify the benefits of GI, Table 4.6 gives the weighting and degree of importance of each benefit. Table 4. 6 Weighting of GI benefits

Benefits Weight Human health and well-being 20% Temperature moderation 17% Water management 15% Community liveability 13% Energy efficiency 12% Biodiversity conservation 12% Visual and aesthetic values 12% Carbon sequestration and storage 10% Pollutant removal 10% Property values 10% water quality improvement 9% Recreational opportunities 8% Noise reduction 7% Food production 5%

Several themes were consistent among all interviewees. The GI approach:  is an imperative for national, regional and local policy regarding sustainable development;  brings economic and health benefits;  contributes to climate change mitigation and adaptation;  can offset the negative environmental and social effects of development; and  improves the quality of life and the quality of place

- GI structure

Interviewees were asked to explain the best scale at which green infrastructure should be applied, and to rate the importance of GI connectivity and networks (Table 4.7). Thirteen participants (62%) believed that GI assessment tools should be applicable at both individual and integrated scales; five (24%) said the tool should be able to assess GI performance as an integrated system. The remaining three (14%) suggested that, due to lack of data, the individual component level is the best scale for assessment.

Table 4. 7 GI structure Integrated components Individual components Both G A P G A P G A P 1 0 4 0 0 3 4 3 6 5 (24%) 3 (14%) 13 (62%) Note: G: Government; A: Academics: P: Practitioners

Section 3: Establishing the framework

Participants were asked whether the proposed framework set out in Figure 2.13 was applicable in the Australian context, or needed to be revised. All interviewees were well versed in the concept of sustainability, sustainable development and Triple Bottom Line analysis of sustainability, yet few were familiar with the concept of Millennium Ecosystem Assessment (MEA) (Table 4.2), although they identified that food, regulation and cultural services were delivered by GI. Interviewees emphasised that the GI framework needs to be benchmarked. Table 4.8 illustrates that most interviewees agreed that benchmarks for each indicator should cover both local and national scales (top-down and bottom-up). This will enable the identification of projects at a local level that can deliver local benefits whilst also contributing to targets at higher levels.

Table 4. 8 Benchmarking scale

Local authority National scale Both G A P G A P G A P 1 0 3 1 1 2 3 2 8 4 (19%) 4 (19%) 13 (62%) Note: G: Government; A: Academics: P: Practitioners

Six of 13 interviewees (46%) who advocated multi-scale benchmarking (Table 4.8) pointed out that establishing the benchmark at an international level is the most consistent way to evaluate the sustainability of green infrastructure projects.

4.3 Stakeholders’ mind mapping

Mind mapping is a diagrammatic technique to gather, organise and interpret information around a particular topic. This map demonstrates how people visualise relationships between various concepts and gives an overall picture of the topic as well as details information that make up the picture (Davies 2011). Figure 4.2 shows the mind mapping of all the participants’ ideas about concept and benefits of green infrastructure, that are organised into a pattern that interprets the relationship between the ideas and their level of importance. The size of the circle indicates the order of important points and less important points. Analysing and coding the interviewees’ responses revealed nine major concepts and themes that were consistent across all interviewees.

Figure 4. 2 Stakeholders’ mind mapping

These nine thematic concepts can be classified into three categories: economic growth; environmental sustainability; and health and well-being (Table 4.9)

 Concept 1: Economic benefits;  Concept 2: Alignment with political issues and city strategies;  Concept 3: Climate change adaptation and mitigation;  Concept 4: Healthy ecosystem;  Concept 5: Biodiversity;  Concept 6: Water management;  Concept 7: Food production;  Concept 8: An active travel network (Physical health aspect);  Concept 9: Enhanced liveability (Social and mental health aspect).

Table 4. 9 GI thematic concepts derived from the interview results

Thematic concepts Categories Coding descriptive statements Increased property values Economic benefits Job productivity Economic growth Alignment with political issues and city Effectiveness of GI implementation at strategies local scale Temperature moderation Wind speed modification Climate change adaptation and mitigation Carbon storage and sequestration Avoided emissions (reduced building energy use) Environmental Air quality sustainability Healthy ecosystem Water quantity and quality Soil structure (increased permeability) Biodiversity Habitat and species diversity Water purification Water management Flow control and flood reduction Food production Provide local food (community gardens) An active travel network Physical well-being Social (social cohesion) well-being Human health and well- (Community liveability and sense of being Enhanced liveability community) Psychological well-being

These nine thematic concepts are inline with green infrastructure framework that was developed by other scholars (Austin 2014; Tzoulas et al. 2007) that was explained in Chapter 2.

4.4 Proposed green infrastructure performance indicators (Initial list)

Green infrastructure performance indicators play an important role in successfully achieving urban sustainability targets. They can be used for proposing new sustainable urban development plans and for improving the decision-making process based on pre-established benchmarks. This will allow the comparison of different practices and facilitate the identification of best practices among various urban development scenarios.

This section proposes a set of initial indicators that links green infrastructure performance into ecosystem services, ecosystem health and human health and well-being. Based on the literature review and results from semi-structured interview with 21 stakeholders in Australia, a set of 30 indictors in four categories, being ecological indicators, health indicators, socio-cultural indicators and economic indicators, are proposed (Table 4.10) that are both qualitative and quantitative. This set of indicators links to the thematic concepts derived from the interviews (Table 4.9) and the GI conceptual framework identified in Figure 2.13 to provide a basis to establish a composite indicator-based model for assessing green infrastructure sustainability performance.

In the next section this initial list of indicators is analysed with input from 373 national and international stakeholders through an online questionnaire to establish an integrated framework by weighting, screening and aggregating selected indicators. This final framework comprises a reduced set of sixteen potential indicators based on experts’ perspectives.

Table 4. 10 proposed green infrastructure performance indicator set Category Indicators Climate and microclimatic modifications (e.g. Urban Heat Island effect mitigation; temperature moderation through evapotranspiration and shading; wind speed modification) Air quality improvement (e.g. Pollutant removal; Avoided emissions) Carbon Emissions (e.g. direct carbon sequestration and storage; avoided greenhouse gas emissions through cooling) Reduced building energy use for heating and cooling (through e.g. shading by trees; covering building by green roof and green walls) Hydrological regulation (e.g. flow control and flood reduction; regulation of water quality; water purification)

Ecological Ecological indicators Improved soil quality and Erosion prevention (e.g. soil fertility; soil stabilisation) Waste decomposition and nutrient cycling Noise level attenuation Biodiversity-protection and enhancement (e.g. Communities; species; genetic resources; habitats) Improving physical well-being (e.g. physical outdoor activity; healthy food; healthy environments)

Improving social well-being (e.g. social interaction; social integration; community cohesion)

Health Health Improving mental well-being (e.g. reduced depression and anxiety; recovery from stress; attention indicators restoration; positive emotions) Food production (e.g. urban agriculture; kitchen gardens; edible landscape and community gardens) Opportunities for recreation, tourism and social interaction (community livability)

Improving pedestrian ways and their connectivity (e.g. increasing safety; quality of path; connectivity and linkage with other modes) Improving accessibility Provision of outdoor sites for education and research

cultural indicators cultural - Reduction of crimes and fear of crime (comfort; amenity and safety)

Socio Attachment to place and sense of belonging (cultural and symbolic value) Enhancing attractiveness of cities (e.g. enhancing desirable views; restricting undesirable views) Increased property values Greater local economic activity (e.g. tourism, recreation, cultural activities) Healthcare cost savings Economic benefits of provision services (e.g. raw materials; timber; food products; biofuels; medicinal products; fresh water etc.)

Value of avoided CO2 emissions and carbon sequestration Value of avoided energy consumption (e.g. reduced demands for cooling and heating) Value of air pollutant removal/avoidance

Economic indicators Value of avoided grey infrastructure design (construction and management costs) Value of reduced flood damage Reducing cost of using private car by increasing walking and cycling (e.g. shifting travel mode)

Summary

The proposed framework that links green infrastructure performance into ecosystem services, ecosystem health and human health and well-being has been tested in this chapter. The framework in Figure 2.13 provides the basis to establish a composite indicator-based model for assessing green infrastructure performance. Semi-structured interviews were conducted to validate this proposed framework. It was demonstrated that it is applicable in the Australian context. Almost all interviewees agreed that sustainability and sustainable urban development are broad ideal concepts that link social, environmental and economic integrity. In addition, all agreed that the triple bottom line is a good concept for developing a framework for GI performance assessment.

However, sustainable development is not the endpoint, it is a transition process that cities undertake moving toward a more sustainable and resilient future (Desha et al. 2016). According to interviewees, a combination of Triple Bottom Line (TBL) of sustainable development and Millennium Ecosystem Services Assessment (MEA) provides the most acceptable approach. The results of this study suggest that practitioners, government agencies and academic researchers are currently considering ways both to introduce GI and evaluate its performance. In the next chapter, potential indicators are derived based on the GI framework, and by conducting an online questionnaire. These indicators will be weighted and aggregated to identify key performance indicators.

CHAPTER 5: ONLINE QUESTIONNAIRE ANALYSIS

Introduction

This chapter presents the methodology used to collect the quantitative data for this phase of the study. A questionnaire survey was issued to 1152 national and international practitioners and academic experts from various sectors of the built environment and sustainability industries. In return, a total of 373 responses were received. The response rate is 32 per cent which is in the range common for such a survey.

The questionnaire was designed to achieve three objectives: 1. Test the hypothesis for green infrastructure definitions 2. Test the hypothesis for green infrastructure framework 3. Identify and select the key indicators from the initial list of 30 indicators to establish a composite index for the next stage of the study.

This chapter comprises five sections. Following this introduction, the first section outlines the design of the questionnaire. The second section presents the survey results including the methodology for the data analysis and the hypothesis testing to establish a framework for the assessment model. The third section analyses the data to identify the relevant and key indicators, which are the result of the participants’ input. In the final stage of analysis, only the indicators that have achieved a score of more than 80% of the weighted average index (WAI) are selected. Lastly, a summary of the methodology and outcomes of the chapter and a list of selected indicators is presented.

A questionnaire was designed and emailed to 1387 selected individuals from representative organisations and the author’s desktop research. These people were selected because they were national and international experts in the field of built environment and/or sustainability. Out of the potential 1387 respondents, 1152 had valid email addresses. The information issued to the respondents included a URL link to the online questionnaire. This information was also distributed to the mailing list of the participant organisations, which were identified in the first round of semi-structured interviews. Additionally, the questionnaire invitation was distributed on the news web page of organisations such as Australian Institute of Landscape Architects, Australian Institute of Architects, Low Carbon Living CRC, Infrastructure Sustainability Council of Australia (ISCA) and Australian Sustainable Built Environment Council (ASBEC). A total of 373 responses were received (32 %).

5.1 Questionnaire design

According to the general rules of a well-constructed survey questionnaire, it must consist of several hierarchical layouts/sections with corresponding objectives for each section. The questionnaire provided to the potential respondents consisted of open and closed-ended questions asking them to choose among various statements that are closest to their own attitude and concerns. A Likert scale was used to present unambiguous positive or negative choices, to which the respondent was asked to indicate agreement on a scale from ‘strongly agree’ to ‘strongly disagree’ (Likert 1967).

The questionnaire contained three sections that included 26 specific questions. ● Section A included seven questions that aim to verify or disprove the definitions of green infrastructure and its performance, and establishes the conceptual framework (Hypothesis testing). ● Section B included nine questions that aim to rate green infrastructure performance indicators and selects the key indicators from among the entire list of indicators. ● Section C comprised ten questions that aim to determine participants’ knowledge and experience.

Table 5.1 explains the questions, their rationale, the measurement scale and the corresponding research objective for the various sections of the questionnaire. The on-line questionnaire can be found in Appendix A.

Table 5. 1 Quantitative Questionnaire Design.

Section Questions and Rationale Measurement Scale Objective Section A: Verify green infrastructure Corresponding questions in this section are derived from Three-point Likert Scale Research objective 1 performance framework detailed literature review, previous studies and results from semi-structured interviews about respondents’ specific attitudes towards: 1. GI definitions 2. GI frameworks / testing three hypotheses (TBL or MEA or combination of both frameworks).

a. Weighting MEA categories Five-point Likert Scale

b. Weighting TBL categories Five-point Likert Scale

Section B: Rating indicators Weighting combination of TBL+MEA categories (derived from Five-point Likert Scale Research objective 1 and 2 semi-structured interviews) Weighting GI performance indicators (30 indicators in four Five-point Likert Scale Research objective 1 and 2 categories based on literature review and results from semi- structured interview) Section C: Expert Classification Data about participants’ background, field of knowledge and Closed and open questions with Research objective 1 and 2 experience various response alternatives

5.1.1 Piloting

A pilot study and preliminary analyses were conducted before the actual data collection commenced to measure the effectiveness of the potential responses and identify vague questions. The questionnaire was pilot-tested with five individual experts. They tested the feasibility and the validity of all questions and the time it took to complete the questionnaire. Following this pilot test, minor changes were made to the research questions: 1. Less paraphrasing and more detailed explanations to eliminate the potential ambiguity for some terms in Section B (indicator descriptions) because several questions were unclear. 2. Shifting the demographic and personal information to section C (near the end of the questionnaire) instead of section A, because for some respondents asking the demographic questions right at the beginning might be seen as confrontational. 3. Changing the five-point Likert scale to a three-point Likert scale in Section A in order to rate proposed definitions and frameworks obtained from the semi-structured interviews and literature review. The reason for doing this was to identify the appropriate definitions and framework very concisely by giving respondents only three options (Agree, Disagree or Neither agree nor disagree). 4. Changing the three-point Likert scale to a five-point Likert scale in Section B for weighting indicators. The reason for doing this was to obtain weightings for all indicators and to help identify the most relevant indicators. The final questionnaire, outlined in Appendix A, consists of three main sections and 26 questions and 60 sub-questions, combining open and closed questions.

5.1.2 Sampling

As indicated by Babbie (1990) and Creswell (2003), sampling is necessary because of the constraints of time which is not possible to study the whole population. Sampling methods involve the selection of a portion of the representative population, which accurately reflects the large entity. Two non-probability sampling methods have been chosen for this study: expert sampling and snowball sampling. Non-probability sampling does not attempt to select a random sample from the population of interest. Rather, subjective methods are used to decide which elements are included in the sample (Kothari 2004).

The expert sampling method is utilised to produce a sample that can be considered representative of the population (Kothari 2004; Tongco 2007). This research requires the assessment and expert opinion of people with a relatively high level of skill and knowledge. The expert sampling method was applied to identify potential respondents with academic and/or practical experience in the field of green infrastructure, sustainable design principles and/or who were involved in developing sustainability indicators and rating tools. Most of the individuals requested to respond to the survey had been identified from their written articles and books, or through projects that were either in the design stage or have already been constructed. 94

In addition, the individuals who received the questionnaire invitation from the news or web page of the selected organisations were encouraged to share the URL link with their colleagues. This technique is called ’snowball sampling’, which is another non-probability method. It is often used in hidden populations that are difficult for researchers to access, or for which a sampling framework is hard to establish and the population size is unknown and (statistically) “infinite”. This method relies on referrals from an initial list of respondents to nominate additional respondents (Goodman 2011; Heckathorn 1997, 2002).

The survey questionnaire was issued to 1387 national and international experts, out of which a population of 1152 valid email addresses was identified. From 1152 requests issued a total of 373 responses were received from the potential respondent population, a response rate of just over 32% (only in relation to the individual emails between 25%-32%). Gay and Diehl (1992) stated that, if the population size is around 500, 50% of the population should be sampled. If the population size is 1,500, 20% should be sampled. Beyond a certain point (at approximately N=5,000), the population size is almost irrelevant, and a sample size of 400 will be adequate (Gay and Diehl 1992, 125). In support of this, Takim et al. (2004) reported that a response rate of 20-30% is a norm in the survey within the construction industry. Therefore, the return rate of 32% out of 1152 representative population size is acceptable.

As mentioned earlier, the population size in this study is statistically infinite because the total number of people who are expert in the field of sustainability, rating tools and green infrastructure at the national and international levels are unknown. Godden (2004) suggested starting with the calculation for infinite populations (Eq. 5.1) and then refining the estimate for finite population (less than 50,000) using Eq. 5.2. The sample size formula for an infinite population is:

SS Z22  p (1  p ) / c Eq. 5. 1 

SS = Sample Size for infinite population

Z = Z value (e.g. 1.96 for 95% confidence level) P = population proportion (expressed as decimal and assumed to be 0.5 or 50%) c = Margin of Error at 5% (0.05)

Based on this formula, the survey would require 384 respondents to achieve a confidence interval of 95% with a margin of ± 5% and population proportion of 50% (the worst case scenario) to be representative of the infinite population.

Using Godden (2004) formula for refining the estimate of sample size for a finite population of 1152 gives 289 participants representative of the population. This indicates that the 373 respondents in this study reaches the minimum number of required sample size.

The sample size correction for finite populations (Godden 2004) is:

New SS/ 1  (( SS  1)/ pop ) Eq. 5. 2 SS pop = population

5.2 Questionnaire findings

The data from the questionnaire were analysed using SPSS 22 (Statistical Package for the Social Sciences). Each question was coded based on either a three-point Likert scale for Section A responses or a five-point Likert scale for the responses for Section B. The values of the three-point Likert scale were coded with: - 1 for ‘Disagree’ - 2 for ‘Neither agree nor disagree’ and - 3 for ‘Agree’. The values of the five-point Likert scale were coded with: - 1 for ‘not important’ - 2 for ‘slightly important’ - 3 for ‘moderately important’ - 4 for ‘very important’ and - 5 for ‘most important’. For the ‘Yes’ and ‘No’ questions: - 1 for ‘Yes’ and - 2 for ‘No’. Although using the mean values for describing the Likert scales is common practice, the issue of treating ordinal scales as interval scales is controversial (Jamieson 2004). Consequently, the minimum, maximum, mode and median value were calculated. Appendix B presents descriptive statistics for each question in more detail.

The following sections present and discuss the findings in detail. The respective analysis methods are also explained and justified.

5.2.1 Participants profile (Results from part C of the questionnaire)

Out of the 373 completed questionnaires, 241 participants were from Australia, or 64.6%. Table 5.2 documents the distribution of the participants in Australia. Of the total, the majority are from New South Wales with 107 or 44.4%. 58 (24 %) are from Victoria, 28 (11.6%) are from South Australia, 25 (10.3%) are from Queensland, 19 (7.8%) are from Western Australia. The remaining participants are equally from Tasmania and Northern Territory (1.9%) (Table 5.2).

Table 5. 2 Distribution of Australian participants

NSW Victoria South Queensland Western Tasmania Northern Australia Australia Territory 107 58 28 25 19 2 2 44% 24% 12% 10% 8% 1% 1%

Table 5.3 shows the distribution of the international participants. This category of participants is divided into six groups based on continent of origin. The majority of international respondents were from Europe (36.5%) with North America a close second (28%). This classification into continents is useful for future studies to help develop composite indicators based on experts’ opinions in different regions around the world.

Table 5. 3 Distribution of International participants

Asia Africa North America South America Europe Oceania

India (8) Nigeria (2) United States (33) Brazil (4) Germany (10) New Zealand (6) China (6) Canada (2) Colombia (1) United Kingdom Singapore (5) Mexico (2) Argentina (1) (8) Turkey (5) Portugal (4) Korea (2) France (4) Iran (2) Netherlands (3) Russia (1) Sweden (3) Saudi Arabia (1) Finland (2) Indonesia (1) Italy (2) Qatar (1) Spain (2) Taiwan (1) Hungary (2) Belgium (2) Iceland (1) Croatia (1) Ireland (1) Lithuania (1) Switzerland (1) Luxembourg (1)

33(25%) 2(1.5%) 37(28%) 6(4.5%) 48(36.5%) 6(4.5%)

In relation to the expertise or background of the respondents, the following was recorded: - 145 (35.63%) are practitioners, - 167 (41.03%) are academics, and - 95 (23.34%) are from state and local government agencies (Table 5.4). Among the Australian participants, eight work in academic and practice divisions, four in practice and government divisions, one in academic and government and three of them in all three sectors. At the international level, seven participants work in practice and academic sectors, one in practice and government sectors, three in academic and government sectors and two in all three sectors.

Table 5. 4 Participants classifications based on job sectors

Sectors Australian International Total Percentage Practitioners 111 34 145 35.63% Academics 65 102 167 41.03% Government 84 11 95 23.34% Total 373 100%

The majority of experts are landscape architects, urban planners, architects and ecologists (Table 5.5). They are involved equally in design/planning and research fields (Figure 5.1 and 5.2). In addition, 60% of the participants have more than 10-years’ work experience (Table 5.6).

Figure 5. 1 Participants’ industry sectors

Table 5. 5 Disaggregation of participants by industry sector

Building/ Infrastructure Landscape Urban Arboriculture/ Ecology Economics Architecture Planning Horticulture 98(26.27%) 75(20.11%) 111(29.76%) 105(28.15%) 24(6.43%) 80(21.45%) 13(3.49%)

Figure 5. 2 Participants’ background and experience

Table 5. 6 Number of years’ work experience

Less than 2 years 2-5 years 5-10 years 10-20 years More than 20 years 20 (5.36%) 52 (13.94%) 82 (21.98%) 117 (31.37%) 102 (27.35%)

In terms of the experience of the experts, it is best described as broad ranging. This is important for the study as it has assisted in collecting a wide sample of knowledge and data across the entire research area (Figures 5.1, 5.2 and 5.3). Figure 5.3 presents the variety of sectors that the participants have experience in. It reinforces the eligibility of participants to take part in this study, and the quality and reliability of the data that is presented based on the participants’ knowledge and experience. A majority of the experts (175 or 46.9%) have been involved in the field of Climate Change and Water Management (159 or 42.6%), Ecology (154 or 41.3%), Energy (149 or 40.00%) and Health and Well-being (114 or 30.6%).

Figure 5. 3 Participants' specific field and focus areas

The following section discusses the results of the first part of the questionnaire, including selection of the relevant indicators and subsequent determination of the important indicators for sustainability measurement based on participants’ responses.

5.2.2 Results from section A of the questionnaire

In section A of the questionnaire, the respondents were asked to provide ‘Yes’ or ’No’ answers to three questions. The questions were focused on identifying the participants’ backgrounds and their familiarity with the three concepts of: Green Infrastructure planning, Triple Bottom Line (TBL) of Sustainability, and the Millennium Ecosystem (service) Assessment framework (MEA). It was found that 359 (96.3%) of 373 experts are familiar with the term ‘Green

Infrastructure’. 287 (76.9%) are familiar with the concept of ‘TBL of Sustainability’ and only 131 (35.1%) are aware of the MEA framework (Table 5.7). It is worthwhile adding that 142 (38.1%) of the 373 participants have been involved directly in the development of sustainability assessment tools, models and frameworks through the contribution of their knowledge to programs or by assisting in the validation of tools.

Table 5. 7 Section A of the questionnaire

Questions Australia International Total Q1. Familiar with the term “Green Infrastructure” Yes 232 (96.27%) 127(96.21%) 359 (96.25%) No 9 (3.73%) 5 (3.79%) 14 (3.75%) Total 241 132 373 (100%) Q2. Familiar with the term “Triple-Bottom-Line” of the sustainability concept Yes 212 (87.97%) 75 (56.82%) 287 (76.94%) No 29 (12.03%) 57 (43.18%) 86 (23.06%) Total 241 132 373 (100%) Q3. Familiar with concept of “Millennium Ecosystem Assessment” Yes 59 (24.48%) 72 (54.54%) 131 (35.12%) No 182 (75.52%) 60 (45.46%) 242 (64.88%) Total 241 132 373 (100%)

In question four, the respondents were asked to note their levels of agreement on four proposed Green Infrastructure definitions using a three-point Likert scale. These definitions have been identified from the semi-structured interviews with the 21 experts discussed in Chapter 4 (Pakzad and Osmond 2015). The four definitions are: Definition 1: Green infrastructure (GI) is a policy and strategic approach to land and species conservation. Definition 2: GI is a network of energy, materials and species flows that maintains and improves ecological functions in combination with multifunctional land uses and provides associated benefits to human populations and ecosystems. Definition 3: GI refers to the integration of ecological functions through natural and engineered networks into conventional infrastructure systems to enhance their functions, and it can significantly reduce their carbon footprint. Definition 4: GI is an ecological solution underpinned by the concept of ecosystem services to improve the sustainability level of the urban and built environment. It embraces the idea of the Triple Bottom Line – the social, economic and environmental aspects of the urban environment.

The values of the three-point Likert scale were coded with: - 1 for ‘Disagree’ - 2 for ‘Neither agree nor disagree’ and - 3 for ‘Agree’.

100

Figure 5.4 illustrates that most of the respondents agreed with definition 2, 3 and 4 with an agreement percentage of 52.2% (206) for D2, 67.6% (252) for D3 and 76.1% (284) for D4. However, 39.4% (147) of participants disagree with the first definition. The overall results for this question indicate that the fourth followed by the third definition implied the best description of green infrastructure.

Figure 5. 4 Summary of answers given by respondents in Q4

Furthermore, participants were asked to give their own definition if they disagreed or neither agreed/disagreed with the four proposed definitions. 35 responses were received. The following are some of their statements. “GI is the integration of ecological functions into conventional infrastructure solutions to improve the sustainability of the urban and built environment.” “GI is the natural assets that support communities.” “I define Green Infrastructure as an umbrella term that includes Water Sensitive Urban Design, Living Architecture (green roofs, walls and facades), Integrated Water Cycle Management, Green streets, Urban Food, and the Urban Forest. These elements form an interrelated web of benefits that combine to make our cities more liveable. These benefits accrue to the general public, private individuals and companies, the greater ecological good, as well as hard and soft economic benefits.” “The concise definition developed by the Green Infrastructure Project in South Australia defines Green Infrastructure as the connected network of green spaces and water systems that delivers multiple environmental, social and economic values and services to urban communities. This definition brings together the ecosystem services delivered by nature and natural cycles, with the linking of green spaces, nature corridors and water networks with specialized forms of engineering infrastructure that replace or add to conventional engineering with ‘green’ elements such as green roofs, living walls and WSUD installations.” “I see GI as a combination of natural and man-made elements in the landscape which collectively contribute to the achievement of a more sustainable way of life - the key challenge 101 is what measures are used to evaluate the performance of GI and how these are then used to inform (and influence) more conventional economic or social performance measures and to make the case for increased investment in GI.” “EU definition: a strategically planned network of natural and semi-natural areas with other environmental features designed and managed to deliver a wide range of ecosystem services. It incorporates green spaces (or blue if aquatic ecosystems are concerned) and other physical features in terrestrial (including coastal) and marine areas. On land, GI is present in rural and urban settings.” “Add to # D4: GI enhances human experience of the built environment by proactively promoting health, beauty, enjoyment of life, interconnectedness of communities, and equity of access for each community member, & optimising the human potential of each individual.”

Table 5.8 shows the distribution of definitions in each job sector. Definitions 1, 2 and 4 achieved a high level of agreement from academic experts. However, practitioners reached agreement on the second definition (D2) in contrast to the two other sectors. In the fourth definition (D4), green infrastructure was defined as an ecological solution underpinned by the concept of ecosystem services that embraces the idea of the ‘Triple Bottom Line of Sustainability’. D4 refers to the combination of ecosystem services and TBL frameworks. This definition is a basis to develop the green infrastructure assessment framework to identify the relevant indicators. To validate this statement in Question 5 (the next question), the participants were asked to note their levels of agreement using a three-point Likert scale for three statements concerning the green infrastructure framework.

Table 5. 8 Descriptive statistics relating to definitions and job sectors

Definition 1: Green infrastructure (GI) is a policy and strategic approach to land and species conservation. Three-Point Likert scale Practitioner Academic Government Total 1.00 147 64(43.5%) 56 (38.1%) 42 (28.6%) (Disagree) (100.0%) 2.00 D1 36 (30.3%) 58 (48.7%) 32 (26.9%) 119 (100.0%) (Neither agree or disagree) 3.00 45 (42.1%) 53 (49.5%) 21 (19.6%) 107 (100.0%) (Agree) Total 145 (38.9%) 167(44.8%) 95 (25.5%) 373 (100.0%) % of Total Definition 2: GI is a network of energy, materials and species flows that maintains and improves ecological functions in combination with multifunctional land uses and provides associated benefits to human populations and ecosystems. Three-Point Likert scale Practitioner Academic Government Total 1.00 17 (44.7%) 11 (28.9%) 14 (36.8%) 38(100.0%) (Disagree) 2.00 D2 54 (41.9%) 48 (37.2%) 38 (29.5%) 129 (100.0%) (Neither agree or disagree) 3.00 74 (64.1%) 108 (52.4%) 43 (20.9%) 206 (100.0%) (Agree) Total 145 (38.9%) 167 (44.8%) 95 (25.5%) 373 (100.0%) % of Total Definition 3: GI refers to the integration of ecological functions through natural and engineered networks into conventional infrastructure systems to enhance their functions, and it can significantly reduce their carbon footprint. Three-Point Likert scale Practitioner Academic Government Total 1.00 13 (46.4%) 12 (42.9%) 7 (25.0%) 28(100.0%) (Disagree) D3 2.00 37 (39.8%) 49 (52.7%) 18 (19.4%) 93 (100.0%) (Neither agree or disagree) 102

3.00 95 (37.7%) 106 (42.1%) 70 (27.8%) 252 (100.0%) (Agree) Total 145 (38.9%) 167 (44.8%) 95 (25.5%) 373 (100.0%) % of Total Definition 4: GI is an ecological solution underpinned by the concept of ecosystem services to improve the sustainability level of the urban and built environment. It embraces the idea of the triple bottom line – the social, economic and environmental aspects of the urban environment. Three-Point Likert scale Practitioner Academic Government Total 1.00 4 (22.2%) 6 (33.3%) 18(100.0%) (Disagree) 9 (50.0%) 2.00 D4 24 (33.8%) 31 (43.7%) 21 (29.6%) 71 (100.0%) (Neither agree or disagree) 3.00 112 (39.4%) 132 (46.5%) 68 (23.9%) 284 (100.0%) (Agree) Total 145 (38.9%) 167 (44.8%) 95 (25.5%) 373 (100.0%) % of Total

Based on the findings from the interviews and literature review, three assessment frameworks were developed (Pakzad and Osmond 2015, 2016). In Question 5, a set of hypotheses was utilised for each proposed framework and participants were asked to determine the level of their agreement with these three hypotheses. They are as follows: HaF1: The TBL Sustainability Framework (Environmental, Social and Economic) can be an appropriate framework for measuring GI performance. HaF2: The Millennium Ecosystem Assessment framework (Provisioning Services, Regulating Services, and Cultural services, Supporting Services + Biodiversity) can be an appropriate framework for measuring GI performance. HaF3: A combination of both frameworks (TBL and MEA) can be an appropriate framework for measuring GI performance.

Analysing the results shows the third framework (Combination of TBL and MEA); followed by the first framework (TBL) achieved the highest scores. Results from the earlier semi- structured interviews indicated a similar outcome.

Figure 5. 5 Identifying the appropriate framework referring to question 5 103

The respondents then were asked to rate the degree of importance of sub-categories of MEA and TBL separately by using a five-point Likert scale. Table 5.9 presents the results for weighting and ranking the subcategories of the MEA framework. ‘Regulating services’ scored the highest. However, all 4 subcategories, with only minor differences, (1-2%), achieved similar levels of importance. The ‘Environmental’ subcategory also achieved the highest score in the TBL framework.

Table 5. 9 Weighting sub-categories of the MEA framework

Millennium Ecosystem 1 2 3 4 5 Mean Weight Rank Assessment sub-categories 51 2 13 119 188 Provisioning services (13.67% 4.28 24.96% 3 (0.54%) (3.49%) (31.9%) (50.4%) ) 259 1 7 22 84 Regulating services (69.44% 4.59 26.76% 1 (0.27%) (1.88%) (5.9%) 22.52%) ) 73 137 130 2 31 Cultural/social services (19.57% (36.73% (34.85% 3.97 23.15% 4 (0.54%) (8.31%) ) ) ) 43 199 4 14 113 Supporting/habitat services (11.53% (53.35% 4.31 25.13% 2 (1.07%) (3.75%) (30.3%) ) )

Total 17.15 100.0 % Statistical analysis based on 373 participants

Table 5. 10 Weighting sub-categories of the TBL framework

TBL sub-categories 1 2 3 4 5 Mean Weight Rank

46 317 1 0 9 36.718 Environmental (12.33% (84.99% 4.82 1 (0.27%) (0.00%) (2.41%) % ) ) 51 149 169 1 3 32.714 Social / cultural (13.67% (39.95% (45.31% 4.29 2 (0.27%) (0.80%) % ) ) ) 76 157 121 3 16 30.568 Economic (20.38% (42.09% (32.44% 4.01 3 (0.80%) (4.29%) % ) ) ) Total 13.12 100.0 % Statistical analysis based on 373 participants

5.2.3 Results from section B of the questionnaire

Based on the results from the semi-structured interviews (Chapter 3) and the literature review, thirty indicators were identified (Chapter 4). These indicators were classified into four main categories (Table 5.11): 1- Ecological indicators 2- Human health indicators 3- Socio-cultural indicators 4- Economic indicators 104

Table 5.1 demonstrates overlapping the sub-indicators of three frameworks , TBL of sustainable development, Millennium Ecosystem Assessment and Green infrastructure assessment framework.

Table 5. 11 Three frameworks

Millennium Ecosystem GI performance assessment TBL sub-categories Assessment sub-categories framework

Ecological indicators Economic Regulating services (ecosystem health)

Environmental Supporting/habitat services

Economic

indicators

Provisioning services Socio-cultural indicators

Social / cultural Cultural/social services Human health indicators

The objective of this section of the questionnaire was to identify the most appropriate indicators for the assessment model. Section B comprises two stages: in the first stage, each participant was given 100 points to distribute over these four subcategories. This approach is the budget allocation technique (BAL). The main advantage of BAL is it is transparent and relatively straightforward to get stakeholders’ opinions on each indicator (Nardo et al. 2005). Table 5.12 gives the average assigned weight of BAL for each category.

Table 5. 12 Weighting main categories result

GI performance assessment framework Mean Number of indicators Ecological indicators (ecosystem health) 32.4% 9 Human health indicators 25.9% 3 Socio-cultural indicators 20.9% 8 Economic indicators 20.8% 10

In the second stage, each participant was asked to score the importance of each indicator in terms of its contribution to sustainability assessment using a five-point Likert scale.

5.3 Analysis to identify the important indicators in measuring the level of sustainability

Statistical analysis of the participants’ responses was undertaken. A descriptive analysis and Cronbach’s alpha reliability test were conducted to identify the central tendencies of the data (see Appendix B). The Cronbach’s alpha result (α=0.93) was higher than the accepted reliability threshold stated by George and Mallery (2011). Afterwards, exploratory analysis was undertaken to evaluate the importance of each indicator within the above four categories and determine the level of agreement among participants by applying the Weighted Average Index (WAI) method (Dutt et al. 2015; Likert 1967). The reason for using the WAI is to identify and select key performance indicators with the highest level of contribution in the assessment

105 model by relying on expert judgment. The WAI for each indicator was calculated by adding up the response numbers for each indicator multiplied by a weighted value between 0.2 and 1 and dividing the sum by the number of total responses. This provides an overall weighted average score for each particular indicator.

Eq. 5. 3  fwii WAI   fi fi = frequency of the respondents wi = the weight of each of the Likert score values assigned as follows:

Not Important (1) 0.2 Slightly Important (2) 0.4 Moderately Important (3) 0.6 Important (4) 0.8 Very Important (5) 1

An example of WAI calculation for the climate and microclimate modifications indicator: WAI= (2*0.2+4*0.4+39*0.6+130*0.8+198*1)/373= 0.878

As listed in Table 5.13, the mean response rating value for the 30 performance indicators offered to respondents ranges from a maximum of 4.53 (improving physical well-being) to a minimum of 3.23 (noise level attenuation). No indicator mean value score fell under the ‘not important’ value (or <1.5). In addition, results of the WAI percentage indicate that the WAI value varies between 64.6% and 90.7%. This means that all indicators are rated above the mid- point of the Likert scale, which implies that all 30 indicators are considered important and can be used to monitor green infrastructure project performance.

Table 5. 13 Indicators included in questionnaire and weighting indicators

Std. WAI Indicators Mean Rank Deviation % Climate and microclimatic modifications (e.g. urban heat island effect mitigation; temperature moderation through 4.39 0.76 87.8% 2 evapotranspiration and shading; wind speed modification) Air quality improvement (e.g. pollutant removal; avoided 4.31 0.75 86.1% 5 emissions) Carbon emissions (e.g. direct carbon sequestration and 4.27 0.86 85.4% 8 storage; avoided greenhouse gas emissions through cooling) Reduced building energy use for heating and cooling (through e.g. shading by trees; covering building by green roof and green 4.28 0.79 85.5% 7 walls)

Ecological Ecological indicators Hydrological regulation (e.g. flow control and flood 4.31 0.77 86.2% 4 reduction; regulation of water quality; water purification) Improved soil quality and erosion prevention (e.g. soil 3.94 0.89 78.8% 22 fertility; soil stabilisation)

106

Waste decomposition and nutrient cycling 3.80 0.93 76.0% 25 Noise level attenuation 3.23 0.96 64.6% 30 Biodiversity protection and enhancement (e.g. communities; 4.22 0.89 84.4% 9 species; genetic resources; habitats) Improving physical well-being (e.g. physical outdoor activity; 4.53 0.69 90.7% 1 healthy food; healthy environments) Improving social well-being (e.g. social interaction; social 4.21 0.84 84.1% 10 integration; community cohesion) Improving mental well-being (e.g. reduced depression and anxiety; recovery from stress; attention restoration; positive 4.33 0.80 86.6% 3

Health Health indicators emotions) Food production (e.g. urban agriculture; kitchen gardens; 4.03 0.93 80.7% 14 edible landscape and community gardens) Opportunities for recreation, tourism and social interaction 4.01 0.81 80.2% 15 (community livability)

Improving pedestrian ways and their connectivity (e.g. increasing safety; quality of path; connectivity and linkage with 4.19 0.87 83.8% 11 other modes) Improving accessibility 3.98 0.91 79.6% 17 Provision of outdoor sites for education and research 3.64 0.96 72.8% 27

cultural indicators cultural - Reduction of crimes and fear of crime (comfort; amenity and 3.63 1.01 72.7% 28 safety)

Socio Attachment to place and sense of belonging (cultural and 3.96 0.93 79.2% 18 symbolic value) Enhancing attractiveness of cities (e.g. enhancing desirable 3.86 0.94 77.3% 23 views; restricting undesirable views) Increased property values 3.26 1.14 65.0% 29 Greater local economic activity (e.g. tourism, recreation, 3.85 0.95 77.0% 24 cultural activities) Healthcare cost savings 3.95 0.99 79.0% 19 Economic benefits of provision services (e.g. raw materials;

timber; food products; biofuels; medicinal products; fresh water 3.78 0.99 76.0% 26 etc.)

Value of avoided CO2 emissions and carbon sequestration 4.03 0.98 81.0% 13 Value of avoided energy consumption (e.g. reduced demands 4.28 0.85 86.0% 6 for cooling and heating)

Economic indicators Value of air pollutant removal/avoidance 4.10 0.88 82.0% 12 Value of avoided grey infrastructure design (construction 3.94 0.95 79.0% 20 and management costs) Value of reduced flood damage 3.96 0.96 79.0% 21 Reducing cost of using private car by increasing walking 4.01 0.97 80.0% 16 and cycling (e.g. shifting travel modes)

107

5.4 Determining key indicators

Reduction of the indicator set to a more manageable number of indicators requires setting a cut-off point. Determining the cut-off point in a Likert scale is very subjective and depends on the aims of the research and how the findings are to be applied. There are several techniques for determining cut-off points: (a) consensus is based on cut-off points at 66.7%, 75%, 80% or 100% agreement (Basinger et al. 2009; Boyd 2003; Budruk and Phillips 2011; Dobbins 1999; Harrison 2005; Pulcini et al. 2006; Teriman 2012; Tigelaar et al. 2004); (b) interquartile range (Beattie et al. 2004; Monette et al. 2013), (c) standard deviation (Brill et al. 2006; Scott 2002; Seibert 2004); and (d) group mean value (Brown et al. 2006).

Amongst all cited techniques, this research used the interquartile range as the effective cut-off point for selecting the indicators to be included in the assessment model. The upper quartile (Q3) is the cut-off point in a distribution above which the highest 25 percent of the WAI scores are located, and the lower quartile (Q1) is the cut-off point below which the lowest 25 percent of the scores are located. (Q2) is usually the median value; it is the value in the data set where there is an equal probability of falling above or below it. Based on the WAI in Table 4, Q1 is 77.3%, Q2 is 80% and Q3 is 85.4%. In this study Q2 or the median has been considered the cut-off point, which delivers a set of 16 indicators achieving 80% and higher on the WAI percentage scale (Table 5.14).

Additional analysis has been undertaken on the final list of key indicators to identify how the different sectors (government, academics and practitioners) scored specific indicators (Table 5.14). It indicates that most ecological indicators gained higher levels of agreement from academics. Health and sociocultural indicators achieved more attention from practitioners and government sectors, and government sector workers favoured economic indicators.

Summary

This chapter has presented the results from a national and international survey to establish an indicator-based framework to evaluate the level of sustainability of GI projects. This initial framework utilises four categories: ecological, social, health and economics, and consists of 30 indicators derived from existing literature, the DPSIR framework and semi-structured interviews (nine thematic concepts). The Weighted Average Index (WAI) method was applied in subsequent analysis to identify the degree of importance of each indicator. It is a quick technique to identify differences in opinions among respondents, and was utilised to select key indicators out of a set of 30.

Bell and Morse (2008) suggest that 20 indicators are manageable for any study, whereas Moles et al. (2008) suggest that up to 40 indicators can be used if time and resources are available. In this study, the selection of indicators was based on a cut-off point of interquartile range (median), which selected all indicators above the median (50th percentile or WAI%≥ 80%). This resulted in selection of 16 indicators, which were rated as either important (score of 4) or 108 very important (score 5). Having achieved more than 80% agreement regarding importance, these 16 indicators were identified as relevant performance indicators for measuring the sustainability level of GI projects.

They derive from all four categories; three health indicators, six ecological indicators, three social/cultural and four economic. These results demonstrate that stakeholders pay greater attention to human health and ecological aspects of GI performance (Table 5.14). ‘Improving Physical Well-being’ and ‘Microclimate Modifications’ were identified as the most important indicators and ‘Reducing cost of using private car by increasing walking and cycling’ as the least important.

Table 5. 14 Key indicators set Leading Indicators WAI% Rank sector Climate and microclimatic modifications 87.80% 2 G/A Air quality improvement 86.10% 5 G Ecological Carbon offset 85.40% 8 A indicators Reduced building energy use for heating and cooling 85.50% 7 A (32.43%) Hydrological regulation 86.20% 4 A Biodiversity-protection and enhancement 84.40% 9 P Health Improving physical well-being 90.70% 1 A/P indicators Improving social well-being 84.10% 10 P/G (25.91%) Improving mental well-being 86.60% 3 G Socio- Food production 80.70% 14 G cultural Opportunities for recreation, tourism and social 80.20% 15 G indicators interaction (20.86%) Improving pedestrian ways and their connectivity 83.80% 11 P Value of avoided CO2 emissions and carbon 81.00% 13 G sequestration Economic Value of avoided energy consumption 86.00% 6 G/A indicators Value of air pollutant removal/avoidance 82.00% 12 G/A (20.8%) Reducing cost of using private car by increasing 80.00% 16 A walking and cycling G: Government A: Academics P: Practitioners

In the semi-structured interview process, when participants were asked to name the benefits of GI, the frequency analysis revealed that human health and well-being, temperature moderation, water management, community livability and energy efficiency achieved the highest levels of importance among 14 classified benefits (Figure 4.1). This indicates consistency of results between two different methods for selecting key indicators.

The final list of key indicators was analysed to identify how different sectors (for example, within government, academics and practitioners) viewed the specific indicators (Table 5.14). Most ecological indicators had the highest level of importance among academics; health and socio-cultural indicators were most important to practitioners and government sectors, while economic indicators was favoured by staff from the government sector (for additional details refer to Pakzad et al. (2017)). In the next chapter, description of each indicator, the parameters and equations used to measure the individual indicators with reference to the scientific literature are presented in more detail. 109

CHAPTER 6: DEVELOPING GIS ASSESSMENT METRICS

Introduction

Analysis from online questionnaires has identified 16 indicators across four categories (ecological, health, sociocultural and economic) which are important for measuring the sustainability performance of green infrastructure (GI) based on experts’ opinions. This set of indicators has then been applied to a case study in the Parramatta CBD area in Sydney, Australia, for further validation and verification of the indicator-based model. The main purpose in developing this indicator-based framework is to make it user-friendly, straightforward, and with less input data required, so that it can easily be used by urban designers, landscape architects and decision-makers. The equations and parameters aim to simplify the complexity of the GI performance measurement system. Each parameter is distinctly correlated and linked to particular physical characteristics and traits of GI types and their context (place to plant and grow), especially street tree canopies.

This chapter is organised into five sections. Following this introduction, the first section gives a description of the equation and baseline value and improvement strategy for each indicator and sub-indicator. Section 2 provides a description of the case study that will serve as a foundation for developing the measurement matrix and for identifying weaknesses and limitations of the model. Section 3 presents the results for the 16 indicators in the study area. Section 4 outlines the aggregation of the results. The chapter concludes with a summary and discussion.

6.1 Description of selected indicators

As discussed in Chapter 2, GI systems of various scales and types can embody a dual rights- based approach, both anthropocentric and ecocentric in value. Therefore, in this research the fundamental approach to estimating the value of GI performance is based on the dual rights concept (humans’ rights and nature’s rights) by considering benefits that humans obtain from urban GI development as well as benefits that ecosystems and surrounding natural environments obtain.

In urban areas, due to conflict with existing buildings and ‘grey’ infrastructure, the quantity and spatial configuration of greenery is often affected by available space and requirements for both root zones and spacing between tree canopies and other GI types. Therefore, benefits, function, structure and distribution of GI are dictated by urban form and structure.

To estimate urban GI values, several relationships must be established. First, it is important to understand the link between existing structures, types and amounts of urban GI with respect to

110 urban form and structure as well as to associated ecological processes. 'Structure' refers to the spatial distribution of vegetation in relation to other objects such as buildings and infrastructure (McPherson et al. 1997; Nowak et al. 2008; Rowntree 1984). Second, it is crucial to understand how changes in the structure of GI due to urban development affect ecological processes and functions. Therefore, careful consideration must be taken when choosing the individual type of GI to be used with respect to specific physical characteristics that will deliver the desired outcomes (e.g. temperature, shade, transpiration etc.) while minimising overall negative environmental and health impacts of urban growth.

Therefore, this indicator-based model assesses GI performance by applying 16 proposed indicators within four categories (ecological, health, sociocultural and economic) in two stages. The first stage involves reviewing the value of existing GI in the study area and the second stage appraises future design scenarios against established baselines for each individual indicator. ‘Baseline’ in this context means the desired sustainability level, which is a subjective and adjustable value, and can be configured for specific urban design targets.

This section presents a description of each indicator followed by the parameters used to measure it. Certain selected indicators have different units and scales that cannot be aggregated in the conventional way by summing up the original results. Therefore, to normalise them, establishing baseline values ranging between 1 and 5 has been proposed (Nardo et al. 2005). They represent maximum and minimum values and refer to different contribution levels of sustainable green infrastructure for each indicator (Table 6.1). These baseline values have been assigned based on the results of other studies that can be applicable in Australia. In some cases, the baseline value was proposed specifically based on the study area and can be adjusted in accordance with the project objectives and desired targets set by stakeholders and decision- makers.

Table 6. 1 Normalisation of indicators by setting baseline values

Level of sustainability Baseline value High 5 Medium-high 4 Medium 3 Medium-low 2 Low 1

6.2 Ecological indicators

Six indicators have been selected to measure the ecological value of green infrastructure (Table 6.2). They consist of: (1) climate and microclimatic modifications; (2) air quality improvement (e.g. pollutant removal and emissions avoidance); (3) carbon offsets (carbon storage and sequestration); (4) reduced building energy used for cooling and heating; (5) hydrological regulation (run-off reduction) and; (6) biodiversity protection and enhancement.

111

Table 6. 2 Ecological indicators

GREEN INFRASTRUCTURE ASSESSMENT MATRIX (PERFORMANCE INDEX)

CATEGORY INDICATOR SUB-INDICATOR DESCRIPTION UNIT Reflection of solar radiation of C1-1: different surfaces. Effective surface Estimation: Albedo of different % albedo surfaces multiply by their area percentages Lowering air temperature through evapotranspiration (percentage of water loss through evaporation and transpiration processes from soil and plants) C1-2: Estimation: multiplying Evapotranspiration reference evapotranspiration by % rate sum of the area percentage of each green cover and their associated three adjustment factors (microclimatic, species and density coefficients) and then dividing by the total green cover area. Climatic and microclimatic Preventing and/or directing air modifications (e.g. Urban movement through green Heat Island effect

corridors is calculated by two mitigation; temperature parameters of canyon geometry C1 moderation through and wind direction of the site: evapotranspiration, - aspect ratio of canyon shading and wind speed C1-3: (H/W) and direction modification) Score Ventilation - street orientation N-S

E-W NE-SW NW-SE

ECOLOGICAL ECOLOGICAL INDICATORS - direction of prevailing wind There is a correlation between shading coefficient of tree species and leaf area.

Urban canopy geometry Urban canopy Estimation: iTree Eco was utilised to calculate leaf area C1-4: (LA) for individual GI types. Then, normalised the sum of LA % Shading effect in each grid cell with z_score formulas and then applied Normsdist procedure to distribute the results to an identical range between 0% and 100%. Estimation: iTree Eco was utilised to calculate pollutants removed by individual GI types. C2-1: Then, normalised the sum of C2 Air quality improvement pollutions removed (CO, O3, % Air pollutants removal NO2, SO2 and PM2.5) in each grid cell with z_score formulas and then applied Normsdist procedure to distribute the

112

results to an identical range between 0% and 100%. Estimation: iTree Eco was C3-1: utilised to calculate carbon offset capacity for each tree C3 Carbon offsets % Carbon storage & based on the tree type sequestration (deciduous, evergreen), tree age, tree height and tree health C4-1: Estimation: iTree Eco was Energy saving utilised to calculate energy Reduced building energy saving for cooling. based on: C4 % used for cooling Tree azimuth, height (>=6 m) and distance from building (<=18 m) C5-1: Estimation: Calculating three Water quality water pollutants removal: (TN, TP, TSS) as a result of Hydrological regulation percentage impervious surfaces % (e.g. flow control and flood (runoff coefficient) and C5 reduction; regulation of percentage pollutants removal water quality; water by each GI type purification) C5-2: Estimation: precipitation Avoid surface runoff interception model in iTree Eco % was utilised C6-1: Habitat suitability was Promoting considered as a proxy indicator. conservation (Habitat Estimation: Least-cost analysis connectivity- method was applied based on suitability)- Fauna resistance surface score for five % selected factors including: land- Biodiversity-protection and use/landcover, distance from enhancement (e.g. roads, distance from water C6 Communities; species; bodies, slope and elevation. genetic resources; habitats) C6-2: Diversity of flora species in open Enhancing diversity of green spaces, streets trees as species (Species well as potential role of the diversity)- Flora green backyards and roofs to % enhance biodiversity. Estimation: Shannon’s equitability score

To measure four indicators: carbon storage and sequestration, energy saving, runoff reduction and air quality and also their monetary values, i-Tree Eco v6 - a freely accessible and user- friendly program - was utilised (United States Department of Agriculture Forest Service 2016). i-Tree is a scientific and peer-reviewed model initially developed in the United States by key scholars Nowak and Crane along with the USDA Forest Service to estimate ecosystem services and benefits of structural characteristics of urban and rural forests. This program is currently being adapted for use throughout Australia, Canada and the United Kingdom with preprocessed pollution and weather data available via the application.

Tree measurements, field data, air pollution and meteorological data (if required) are entered into the i-Tree Eco application manually. These input data support the model to calculate structural and functional information using a series of scientific equations or algorithms. A recent version of this application (Version 6) has been modified so that it is capable of estimating the benefits of canopies with less input data so that users have flexible tree

113 measurement options for running the model (United States Department of Agriculture Forest Service 2016). This current version provides an estimate of urban tree structure, pollution reduction, hourly urban tree volatile organic compound emissions, yearly avoided runoff, total carbon stored and annual carbon dioxide sequestration by trees, as well as amount of energy saved by providing shade and optimising ventilation to buildings (and subsequent reduction of carbon dioxide emissions from power plants for heating and cooling energy).

In more detail, the following sub-sections describe the six ecological indicators and sub- indicators followed by the parameters and equations used to measure the individual indicators, units and baseline values.

INDICATOR C1: Climatic and microclimatic modifications

Trees reduce heat stored in horizontal surfaces and facades by providing shade and intercepting incoming solar radiation. Furthermore, their leaves convert solar radiation into latent heat (evapotranspiration), which results in a more moderate air temperature and higher humidity. Therefore, green surfaces and their surroundings are cooler compared to built surfaces and their environments.

A considerable number of studies have been undertaken to investigate outdoor thermal environmental factors and approaches to providing cooling and a thermally comfortable environment in cities. According to Lin et al. (2010), the outdoor thermal environment is significantly influenced by anthropogenic heat (Ichinose et al. 1999); the thermal capacity of urban surface materials (Akbari and Taha 1992; Lin et al. 2007); urban geometry (Arnfield 2003; Oke 2002); evaporation and evapotranspiration rates of plants and wet soil (Dimoudi and Nikolopoulou 2003; Robitu et al. 2006); and the shading effect by trees and other human-made objects on land surfaces (Lindberg and Grimmond 2011).

Two main approaches exist in the climatological assessment of thermal environment conditions for humans: “(1) the physical approach, which is expressed by means of comfort and (2) the adaptive approach, which is subjectively gathered from social survey” (Ali-Toudert et al. 2005 p.246). Most of the field studies that investigate outdoor thermal conditions usually use indoor models and indices that are based on steady-state heat and energy balance equations of the body. They are far from realistic in relation to outdoor thermal conditions and, for example, do not consider wind flow and the additional effects of solar radiation on the outdoor environment.

However, there are various adaptation and mitigation strategies applied to providing a cooling effect in cities. Several studies have used a number of models and standards such as the UC Berkeley thermal comfort model (UCB), ASHRAE Thermal Comfort tool based on Fanger’s PMV/PPD model (Predicted Mean Vote/Predicted Percentage of Dissatisfied) (ASHRAE 2004), the Physiologically Equivalent Temperature (PET), the Outdoor Standard Effective Temperature OUT-SET (Pickup and De Dear 2000) and the Dynamic Thermal Stimulus model (DTS). They are widely recognised to measure outdoor human thermal comfort (Ali-Toudert and Mayer 2007; Cheng et al. 2012). Researchers have applied these models in different 114 climate areas. The overall results from these studies regarding urban design criteria can be very subjective in any given climate, but in general, it has been concluded that low-rise buildings with wide streets (for better ventilation); shade (provided by vegetation); and closeness to a body of water, as well as NE-SW or NW-SE street orientations, all deliver a more thermally acceptable environment, at least in warm climates (Ali-Toudert et al. 2005; Johansson and Emmanuel 2006; Matzarakis et al. 2009; Norton et al. 2015; Jennifer Spagnolo and Richard De Dear 2003; JC Spagnolo and RJ De Dear 2003). In addition to these design considerations, urban greenery provides major benefits to the urban microclimate and thermal comfort (Lindberg and Grimmond 2011) and the overall quality of life of residents (Russo et al. 2016). It modifies surface and air temperatures due to evapotranspiration and shading as well as alters surface albedo (Bowler et al. 2010; Di Leo et al. 2016; Norton et al. 2015).

Not surprisingly, the microclimatic benefits of urban greenery, regardless of type and size, have been well recognised in the literature and in professional practice worldwide (Akbari et al. 2001; Ali-Toudert and Mayer 2007; Bowler et al. 2010; Di Leo et al. 2016; Huang et al. 2008; Russo et al. 2016; Shashua‐Bar et al. 2010). Many local government agencies have set a target of increasing tree canopy in their urban forest strategy in order to mitigate high surface temperatures and the urban heat island effect. For example, the City of Sydney plans to increase vegetation cover from 15.5 to 23.25% by 2030 while the City of Melbourne aims to increase tree canopy cover from 22 to 40% by 2030. The question is: under which circumstances should these projected amounts of vegetation/tree canopies be distributed so as to reach the target value and provide the greatest benefits to pedestrians and urban residents?

Implementing tree canopies and other types of vegetation covers has more limitations in an urban canyon compared with open spaces and urban parks because of available space, street width, building height, street orientation, above- and below-ground grey infrastructure elements and surface characteristics. These factors also influence the capability of a tree to deliver benefits, particularly in microclimatic performance. For example, the potential cooling benefit of trees in a wide canyon with E-W orientation and low aspect ratio is higher than that of a N-S canyon with either high or low aspect ratios (Sanusi et al. 2016). E-W (wide) canyons without trees are the most stressful and uncomfortable in comparison to the other street directions (Ali-Toudert and Mayer 2007). On the other hand, the microclimatic effect of a tree canopy in deep and narrow streets is insignificant because of the self-shading of tall buildings.

As a result, in order to calculate the climatic and microclimatic benefits of urban greenery, four parameters have been investigated: (1) surface albedo, which gives information about the material of the land surfaces (in this study, only horizontal surfaces) and determines the percentage of reflection of solar radiation; (2) evapotranspiration rate, the cooling effect of a vegetated area compared to a non-vegetated area; (3) ventilation - related to street orientation with respect to the prevailing winds and; (4) shading effect - the amount of shade that is provided by buildings, trees and other obstacles on the ground surface (Coutts et al. 2012; Coutts et al. 2016; Norton et al. 2013; Norton et al. 2015).

115

INDICATOR C1.1: Effective surface albedo

Description: The term “albedo” refers to a surface’s ability to reflect incoming solar radiation and is calculated using the following equation:

kk Eq. 6. 1   dir dif k

Where, α represents albedo, kdir represents directly incoming short-wave radiation emitted by the sun, kdif is incoming diffuse short-wave radiation and k↑ represents the total amount of short- wave radiation reflected by the surface. Albedo thus describes the reflectance of a surface ranging from 0, representing the surface absorbing all radiation, to 1, representing it reflecting all radiation (Akbari et al. 1992).

Calculation: Changes in surface albedo affect net radiation and the balance of energy and mass in the planetary boundary layer. There are various techniques for determining surface albedo. However, the most commonly used one is with satellites which carry sensors, for example, TM - Landsat 5, ETM+ - Landsat 7, AVHRR-NOAA, MODIS - Terra/Aqua, ASTER - Terra, among others. The albedo has also been estimated for climate modelling and assessing impacts on air temperature. In these studies, albedo was evaluated by using combinations of reflective satellite bands involving radiation and energy balances based on the method that was initially proposed by Zhong and Yinhai (1988). This technique was subsequently used and developed by other scholars extensively (Allen et al. 2007; Bastiaanssen et al. 1998; Hansen et al. 2003; Román et al. 2013; Wang et al. 2014) .

This technique requires the identification of weights, or the relative contribution of each spectral band, to the composition of the albedo across the entire solar radiation spectrum (Allen et al. 2007). In this study, to identify differences in winter and summer albedo, Landsat 8 images have been used. The specific solar constant (Kb) of each one of the Operational Land Imager (OLI) bands from 2 to 7 has been obtained based on the values of Kb (W m-2 μm-1 sr- 1) for selected days in both summer and winter. As observed for each OLI image, these values were in sharp contrast to each other, resulting in the alterations of Addrad and Multrad, which were used for the radiometric calibration of Landsat 5 TM and Landsat 7 ETM+ images. Based on these results, it is recommended not to use the methodology that employs Addrad and Multrad, but rather to use only the reflectance and planetary albedo with mean weights (as in Zhong and Yinhai's technique).

This technique needs numerous data inputs and the capability to use additional software such as ENVI. ENVI software has capability to analyse and visualise remotely sensed data e.g. LiDAR. This tool can extract information such as land cover from imagery. However, this method has been employed specifically in the case study of Parramatta CBD to validate the range of albedo for each surface type presented in Table 6.3 that were adopted from other studies (Akbari et al. 2001; Akbari and Taha 1992; Christopherson 2007; Conway and The 116

Maryland Space Grant Consortium 1997). The average surface albedo value in each grid cell (100m ×100m) was subsequently calculated by multiplying each surface type by its average albedo value from Table 6.3 and then dividing the sum of these results by their total area, as shown in the formula below:

()A  Eq. 6. 2 i ii ESA 100  Ai

Where ESA is the effective surface albedo ratio in each grid cell (100m x 100m), Ai is area of 2 each surface cover (m ), and αi is the average albedo value of each surface type.

Table 6. 3 The approximate ranges of surface albedos . Sources:(Akbari et al. 2001; Akbari and Taha 1992; Christopherson 2007; Conway and The Maryland Space Grant Consortium 1997). Surface Albedo Average surface Surface cover type range albedo value (α) Urban trees 0.15 – 0.18 0.165 Shrubs/long grass (1.0 m) 0.15-0.30 0.225 Short grass (0.02 m) 0.10-0.25 0.175 Soil 0.10-0.25 0.175 Asphalt 0.05-0.10 0.075 Concrete 0.10-0.35 0.225 Dark roof 0.08 – 0.18 0.13 White roof 0.35 – 0.50 0.425 Water 0.05 – 0.22 0.135

Normalisation: Typically, urban albedos fall in the range of 10 to 20% (0.1-0.2) (Taha 1997). However, in some cities, these values can be exceeded. North African towns are good examples of high albedo urbanised areas (albedos of 30 to 45%), whereas most US American and European cities have lower albedos (15 to 20%). For example, Taha (1994) found that urban areas usually had an albedo of between 12 and 16% and the value of albedo in a CBD with higher percentage of built surfaces was higher than that of its surroundings, partly because of more extensive vegetation in the surrounding areas. Trees and other vegetation have a low (level of) surface albedo as they absorb most of the radiation in the visible and ultraviolet spectrum for the process of photosynthesis. Table 6.4 presents the normalised value of surface albedo. Based on the results from calculations of summer and winter surface albedo in the study area, the albedo range has been adjusted. Therefore, five reference levels were equally assigned between this range (12-28).

Table 6. 4 Normalisation of surface albedo value

Effective surface Albedo (%) Baseline value >28 5 22.6-28 4 17.3-22.6 3 12-17.3 2 <12 1

117

Improvement strategy/Future scenario testing: According to Table 6.4 high albedo materials have a higher baseline value. Therefore, in terms of increasing the reflection rate of materials and finishing surfaces to increase the cooling effect in each grid cell (100m x 100 m) of the study area, the proposed improvement scenario is to use high albedo materials as well as increasing shrubs and canopy covers in each grid cell to reduce the amount of solar radiation absorbed through building envelopes and urban structures and to keep their surfaces cooler.

INDICATOR C1.2: Evapotranspiration rate

Description: Evapotranspiration (ET) is a key component of the hydrological cycle and represents the loss of water through the combined processes of evaporation from soil and vegetation surfaces and transpiration from plant leaves. This process is affected by solar radiation, air temperature, humidity, wind speed and soil moisture (Allen et al. 1998; Burt et al. 2005). Transpiration is the vaporisation of liquid through plants’ stomata during photosynthesis. Plant characteristics (species, leaf area, nutrient level) also have a significant influence.

Evaporation and transpiration happen at the same time and distinguishing between the two processes is difficult. When the plant is small and most of the solar radiation is reaching the soil, close to 100% of ET comes from evaporation through soil surface (Figure 6.1) but over the growing period, once a crop completely covers the soil below it (maximum leaf area per unit surface), transpiration becomes the main process, about 90% of ET (Allen et al. 1998).

Figure 6. 1 The partitioning of evapotranspiration into evaporation and transpiration over the growing period for an annual field crop (Adopted from Allen et al. (1998)).

Calculation: The calculation of ET traditionally was conducted in order to determine water needs of agricultural crops for scheduling and planning irrigation systems (Allen et al. 1998; Liu et al. 2002; Wright 1982). Crop Evapotranspiration (ETC) is calculated by multiplying the reference evapotranspiration (ET0) by the crop coefficient, or species factor (Kc), as presented below:

118

Eq. 6. 3 ETCC ET0 K

-1 Where, ETC is crop evapotranspiration (mm d ), representing the total amount of water lost during a specific period of time; KC is the crop coefficient (dimensionless), which varies based on specific crop characteristics and alters during the crop development stages; and ET0 is reference evapotranspiration (mm d-1). KC is determined from field research and expresses the amount of water that a given species of plant needs during its growing period. ET0 is reference evapotranspiration rate, which represents a regionally specific estimate of the amount of water lost from well-watered grass with 0.12 m height, a fixed surface resistance of 70 sm-1 and an albedo of 0.23, growing in open-field conditions. The value is used as a baseline for water demand for landscaping in the given region and is presented in mm of rain per month. ET0 can be computed from meteorological data using the FAO Penman-Monteith (1965) equation. This equation requires local data such as temperature, solar radiation, and wind speed (to calculate reference evapotranspiration). All these data are available through meteorological stations, most of which have local daily and monthly ET0 available.

900 0.408 (R  G )  u ( e  e ) Eq. 6. 4 nT  273 2 s a ET0    (1  0.34u2 )

-1 Where, ETo = reference evapotranspiration (mm day ), Rn= net radiation at the crop surface (MJ m-2 day-1), G= soil heat flux density (MJ m-2 day-1), T = mean air temperature at 2 m above -1 the ground (°C), u2= wind speed at 2 m above the ground (m s ), es = saturation vapour pressure (kPa), ea = actual vapour pressure (kPa), es - ea = saturation vapour pressure deficit (kPa), Δ = slope of the vapour pressure curve (kPa °C-1), γ = psychrometric constant (kPa °C-1).

Apart from agricultural crops, researchers realised ET needs to be calculated in urban areas with mixed species of vegetation. Currently, remote sensing methods are the most up-to-date and accurate approaches to calculating evapotranspiration in mixed vegetation conditions at extensive spatial and temporal scales. However, special expertise is needed to analyse satellite images, and it comes at a high cost, associated with obtaining high-resolution images.

In addition, there are a number of techniques which are simpler and more straightforward for estimating landscape evapotranspiration (ETL), based on ET0 and landscape coefficient (KL) like the “landscape coefficient method‟ and the “landscape irrigation management program‟ methods proposed by Costello and Jones (1994) and Snyder and Eching (2005) respectively. Yet, despite the broad range of methods and techniques, ET estimations of mixed landscape vegetation in an urban context are typically insufficiently determined due to the complexity of diversity in water needs of mixed species (Costello and Jones 2014; Nouri et al. 2013; Snyder et al. 2015).

119

In this study, the method applied for calculating evapotranspiration in an urban area is the Water Use Classification of Landscape Species (WUCOLS) method developed by Costello and Jones (1994) which is a straightforward and simple technique to calculate ETL (Costello and Jones 2014; Costello et al. 2000). Landscape evapotranspiration (ETL) for urban vegetation is calculated by multiplying the landscape coefficient (KL) by the reference evapotranspiration (ET0) as follows:

ET K ET Eq. 6. 5 LL 0

Values for reference evapotranspiration rate (ET0) can be obtained from Table 6.5 for Australian cities, or, for more accurate results ET0, as mentioned earlier, can be estimated with the Penman-Monteith (PM) equation (Monteith 1965), as proposed by the United Nations FAO Irrigation and Drainage Paper (FAO 56) (Allen et al. 1998).

Table 6. 5 Evapotranspiration Pan Data- Mean daily (mm per day) adopted from (Connellan 2002). Summer Autumn Winter Spring LOCATION Annual (January) (April) (July) (October) Adelaide 8.7 4.3 1.9 5.5 1898 Brisbane 7.4 4.4 3.0 6.5 1960 Canberra 7.0 2.7 1.2 4.0 1375 Darwin 6.7 7.1 7.3 8.6 2696 Hobart 4.9 2.1 0.9 3.1 994 Melbourne 6.1 2.8 1.3 3.8 1268 Perth 8.1 4.0 2.0 5.0 1764 Sydney 7.0 4.1 2.7 5.7 1788

Solving the PM equation requires obtaining meteorological data from the closest weather station to the site. Weather input data includes the daily mean of solar radiation (MJ m-2 d-1), air temperature (oC), wind speed (m s-1), and dew point temperature (oC). If only temperature data is available, then the Hargreaves equation (1982) is used. The Hargreaves equation is simple and is used to estimate solar radiation as well (Hargreaves and Samani 1982, 1985).

Eq. 6. 6 1 TTmax min ET0 a  b  0.0023  17.8  T max  T min  Ra  2

Where, Tmax (Celsius) and Tmin (Celsius) are the maximum and minimum daily air temperatures; Ra (MJ m-2 d-1) is extra-terrestrial solar radiation. The parameters 'a' (mm d-1) and 'b' are calibrated coefficients, determined based on a monthly or yearly basis by regression analysis or visual fitting. An unadjusted version of the Hargreaves equation gives by default a=0 and b=1 (Hargreaves and Samani 1985).

In landscape evapotranspiration formulas, KL is the landscape coefficient that has the same function as the crop coefficient (KC), but it is determined by multiplying three dimensionless adjustment factors: species (Ks), density (Kd), and microclimate (Kmc). 120

K k  k  k Eq. 6. 7 L s d mc

Where, Ks is the species coefficient which reflects the amount of water that a given species of plant needs, Kd is the density coefficient that adjusts the ETL based on vegetation density and configuration, and Kmc is the microclimate coefficient that adjusts the ET0 for the “local” microclimate. Table 6.6 represents the range of value for each landscape coefficient.

Table 6. 6 Landscape coefficients range from very low to high (Costello et al. 2000, 22)

Range Ks Kd Kmc High 0.7-0.9 1.1-1.3 1.1-1.4 Mod/Ave 0.4-0.6 1.0 1.0 Low 0.1-0.3 0.5-0.9 0.5-0.9 Very low <0.1

The species coefficient (ks): This factor is very subjective and ranges from 0.1 to 0.9. It is classified into four categories based on water use studies (Table 6.7), i.e. very low (<0.1), which means plants in this category require water less than 10% of ET0, low (0.1-0.3), moderate (0.4- 0.6), and high (0.7-0.9). In the mixed planting zone with different ranges of water requirements for each species (different KS values), the highest KS value is selected for the calculation. Consequently, for the improvement strategy, substituting the highest Ks species with a lower Ks species should be considered.

Table 6. 7 Categories of water needs (Costello et al. 2000)

Category Percentage of ET0 High 70-90 Moderate 40-60 Low 10-30 Very low <10

Costello and Jones (2014) classified relative water requirements for approximately 3500 landscape plant species in six climatic zones with reference cities in California that can be used as a guide to identify Ks for given species on the site. In addition, the Green Star rating tool listed Ks for number of common species in urban area to help assessors to find the appropriate factor for water requirements of plants (Green Building Council of Australia 2014).

The density coefficient (kd): The density factor refers to the percentage of canopy cover and its distribution in a landscape zone. This factor is selected from one of three categories: low or sparse (0.5–0.9); average or one species (1.0); and high or mixed (1.1–1.3). Immature and sparsely planted landscapes, with less leaf area (less than 60% of the surface coverage in vegetation) are assigned a low category kd value. Plantings with mixtures of trees, shrubs and groundcovers are assigned a density factor value in the high category. Plants that cover 100 121 percent of the planting zone, but are predominantly of one vegetation type, are assigned to the average category. Based on agricultural studies, the water loss from orchards does not increase as canopy cover increases from 70% to 100%. However, below 70% plantation coverage in the planting zone, the trend of water loss is decreasing gradually (Costello et al. 2000). So, the Kd would be considered as average if tree cover is greater than 70% or if groundcover and shrub canopy cover is greater than 90%. Less than 90% shrub canopy cover would be in the low category. Recommended (Kd) values from the Green Star rating tool for less than 40% green coverage are 1.2, 1.0 and 0.8 for dense, normal and sparse planting areas (Green Building Council of Australia 2014).

The microclimate coefficient (kmc): This factor is estimated using three categories that relate local microclimate to regional ET0, ranging from 0.5 to 1.4, i.e., low (0.5–0.9), average (1.0) and high (1.1–1.4). The “low‟ microclimate condition is common when plantings are shaded for most of the day or are protected from strong winds. Based on the LEED rating tool a value of 0.8 is recommended to apply for Kmc (U.S. Green Building Council 2015). An “average‟ microclimate condition is equivalent to reference ET conditions: (Kmc=1.0) which is an open- field setting without extraordinary winds or heat inputs for the location. In rare circumstances where a zone is exposed to very high winds and sun (e.g. planting near street medians, parking lots) or are exposed to reflected surfaces with high albedo that will increase evapotranspiration, a very high microclimate factor (Kmc >1.2 and ≤ 1.4) is applicable. In this situation, the LEED rating tool in the United States recommends to use the highest value of 1.2 for the microclimate factor (Kmc = 1.2) (U.S. Green Building Council 2010) while the Green Star rating tool in Australia suggests Kmc=1.3 is more appropriate. Similarly, where a zone is heavily shaded and protected from wind and sun, a very low microclimate factor (Kmc < 0.8) may be appropriate, the Green Star rating tool recommends Kmc=0.7 and in extreme situations this factor can decrease to 0.5 (Green Building Council of Australia 2014).

Here, the landscape evapotranspiration (ETL) in each grid is calculated and normalised individually as follows.

()AKKK    Gi si di mci Eq. 6. 8 ET ET  100 L 0 A  G i

Where, AGi is the area of green cover in each zone multiplied by density, species and microclimatic factors of given species and divided by the total planting areas in each grid cell.

Estimated values for Ks, Kd and Kmc based on vegetation types in five categories: trees, shrubs, ground cover, turf and mixed planting area can be obtained from Table 6.8.

Table 6. 8 Estimated values for species and density, microclimate (Harris et al. 2004, p.324; U.S. Green Building Council 2015). Vegetation Type Species Factor (Ks) Density Factor (Kd) Microclimate Factor (Kmc) low avg. high low avg. high low avg. high Trees 0.2 0.6 0.9 0.5 1.0 1.2 0.5 1.0 1.4

122

Shrubs 0.2 0.5 0.7 0.5 1.0 1.1 0.5 1.0 1.3 Groundcovers 0.2 0.5 0.7 0.5 1.0 1.1 0.5 1.0 1.2 Mixed trees, 0.2 0.6 1.0 0.6 1.1 1.3 0.5 1.0 1.4 shrubs, groundcovers Turf 0.6 0.7 0.8 0.6 1.0 1.4 0.8 1.0 1.2

Normalisation: Originally, the WUCOLS method was proposed to estimate water loss from landscape plantings (ETL) through evaporation and transpiration for substituting amount of water lost by irrigation (Costello and Jones 1994). Their formula covers three factors: specific on-site weather conditions; plant physiology (structure); and plant combinations (distribution).

It is assumed that species requiring a greater amount of water transpire and evaporate at a higher rate than those needing less water. Species with the higher rate of (ETL) or water loss have significant contribution to microclimatic modification due to increasing moisture of the surrounding area and cooling effects on air temperature. Therefore, according to Costello and Jones (2014) if a species requires 70-90% of the ET0 rate water, then the evapotranspiration coefficient is listed as high; those requiring 40-60% and 10-30% of ET0 rates are listed as moderate and low KL, respectively, and less than 10% of ET0 should be listed with a very low KL which means the species needs less water due to a lower transpiration rate (Table 6.9).

Table 6. 9 Normalisation of evapotranspiration rate (Costello and Jones 2014; Costello et al. 2000).

Evapotranspiration rate Baseline value >70 5 50-70 4 30-50 3 10-30 2 <10 1

Improvement strategy/Future scenario testing: To achieve the highest level of evapotranspiration/moisture in each grid (100 m by 100 m), based on the landscape evapotranspiration formulas (ETL), one or all three factors (Kmc, Kd, Ks) can be altered to achieve the higher rate of transpiration by plants. For example, to achieve the highest level of evapotranspiration, for one of the improvement scenarios, proposing planting with mixed and high density composition with the highest KS value while locating plants in the most windy or sunny setting or close to a high albedo building facade is recommended.

123

INDICATOR C1.3: Ventilation (Blocking and directing air movement)

Description: In the urban context, natural ventilation, or permeability of airflow, that is dictated by urban geometry has a significant impact on microclimate. Urban geometry is often defined as street orientation and the height-to-width (H/W) or aspect ratio of urban street canyons (Figure 6.2): the higher the H/W ratio, the less substantially the trapped heat from the canyon materials is released as a result of reduced sky view factor (SVF), which results in increased air temperatures, especially during the night (Ali-Toudert et al. 2005).

Modelling air movement and ventilation and its climatic impacts in an urban area is a complex task. According to Oke (1988), the significance of urban canyon orientation varies according to the climate zone in a city, it may be influenced by site-specific features such as prevailing winds or topographic conditions. While it may be important in hot climates to minimise solar exposure to buildings, maximise the amount of shade for pedestrians and streets, and facilitate ventilation, cities in cold climates might try to maximise limited sun access, promote protection from cold winds and purposely generate heat islands (Johansson and Emmanuel 2006).

Figure 6. 2 Street orientations and H/W ratio (Ali-toudert and Mayer 2006).

Thermal comfort is defined as “the condition of mind that expresses satisfaction with the thermal environment and is a matter of individual perception” (Standard 2010). Mean radiant temperature, ambient air temperature, relative humidity and wind are four factors that significantly influence the ability of a human body to maintain its core temperature and are thus defining parameters for human outdoor thermal comfort. Table 6.10 provides an overview of the “Beaufort scale”, which is used for the classification of wind speeds and their associated

124 effects on land. Generally, air velocity of more than 5 metres per second (18 km/h) creates discomfort, 10 metres per second (36 km/h) is “definitely unpleasant” and more than 20 metres per second (72 km/h) is considered dangerous (Penwarden and Wise 1975).

Table 6. 10 Beaufort scale (Penwarden and Wise 1975)

Beaufort Number Descriptive Term Speed (km/h) 0 Calm <2 1 Light air 2-5 2 Light Breeze 6-11 3 Gentle Breeze 12-19 4 Moderate winds 20-29 5 Fresh winds 30-39 6 Strong winds 40-50 7 Near gale 51-62 8 Gale 63-75 9 Strong gale 76-87 10 Storm 88-102 11 Violent storm 103-117 12 Hurricane >118

There is a clear link between urban geometry and air temperature. A study conducted by Ahmed (1994) in Dhaka, Bangladesh during the hot and humid summer indicated that air temperature decreased with increased H/W ratios. Likewise, Johansson (2006) found in the hot but dry climate of Fez, Morocco, that canyons with a high aspect ratio had a substantially lower air temperature than a shallow (low H/W ratio) street canyon. Also, another study in Colombo, Sri Lanka, where the climate is hot and humid, revealed maximum daily temperature differences of up to 7 degrees Celsius between sites of different H/W ratios (Johansson and Emmanuel 2006). Conversely, street canyons with high H/W ratios have a negative effect during night times since blockage of the sky view, which is caused by high H/W ratios, reduces heat loss through net long-wave radiation to the sky. The results indicate, however, that the positive effect of shading during daytime is more significant than the negative effect at night.

Calculation: In this study, one of the key principles of urban design in terms of canyon geometry (H/W) and street orientation have been employed to develop wind movement benefit suggestions and establish the baseline value in Sydney.

 H Eq. 6. 9 VEN  canyon& Orientation  W

Where, VEN is ventilation score, H is the average height of tree canopy or building (depending on which one dominates) alongside of the canyon, W is width of canyon.

125

Table 6. 11 Wind direction in Sydney - Principal times of the year for wind occurrence in Sydney (Goal: Direct NE, Prevent S and W)

Month Prevailing Wind Direction North-Easterly Southerly Westerly Jan-March x x April x x May-Aug X Sep x X Oct-Dec x x

Figure 6. 3 Wind rose- Sydney airport (1995-2014)- all months 10m height- Calm- 0.6%

Normalisation: Normalisation values for ventilation based on canyon geometry and street orientation are presented in Table 6.12. According to Ali-Toudert and Mayer (2007), the threshold level for affective aspect ratio H/W is equal to two.

Table 6. 12 Normalisation of ventilation value for Sydney (Source: Author)

Ventilation (Street orientation & aspect ratio) Baseline value NE-SW (along summer wind path) H/W >=2 5 NW-SE H/W >=2 N-S (along secondary wind path) H/W >=2 4 NE-SW (along summer wind path) H/W<2 NW-SE H/W <2 N-S (along secondary wind path) H/W <2 3 E-W (along winter wind path) H/W >=2 2 E-W (along winter wind path) H/W <2 1

Improvement strategy/Future scenario testing: The performance of this indicator is very site-specific. Therefore, to achieve the best natural ventilation in the study area planting trees on the appropriate side of the street is recommended to modify the aspect ratio of the canyon. 126

INDICATOR C1.4: Shading effect

Description: Trees can cool down air temperature not only through evapotranspiration (ET cooling) and ventilation, but also by providing shade to buildings and other heat-absorbing surfaces in the summer and (undesirably) during the winter. This affects outdoor thermal comfort for pedestrians as well as indoor energy consumption required for heating and cooling. Tree structure (tree height, crown size, shape and leaf area), species, orientation, and distance from buildings as well as tree health conditions are key factors for determining the shading coefficients of vegetation (Donovan and Butry 2009; McPherson and Rowntree 1993; Millward et al. 2014; Sawka et al. 2013).

Species shading coefficient refers to the percentage of light intensity intercepted by foliated tree crowns: I/I0, where I is light intensity beneath the canopy and I0 is light intensity above the canopy. It varies from 0.0, where all solar radiation is blocked, to 1.0, where all is transmitted. According to Nowak (1996), in urban areas, the species shading coefficient generally falls within 0.67 to 0.88 (Millward et al. 2014). Determining the shading coefficient is related to the method used to estimate leaf area index (LAI). This method is based on the light intensity under trees in forest stand conditions (densely planted areas) or closed canopy positions (Crown Light Exposure or CLE= 0–1) and uses the Beer-Lambert Law (Nowak et al. 2008) as follows:

LAIln( I / I0 )  I  k Eq. 6. 10

Where, k = light extinction coefficient (Smith et al. 1991). The light extinction coefficient is 0.52 for conifers and 0.65 for hardwoods (Jarvis and Leverenz 1983). In the area with a percentage of planted conifer and/or hardwood trees, k is:

k(% conifers  0.52)  (% hardwoods  0.65) Eq. 6. 11

Calculations: LAI is a dimensionless variable and is defined as the m2 leaf area per m2 projected ground area of a canopy (Gower and Norman 1991). It is often used to estimate carbon flux, evapotranspiration, modelling growth rate, primary vegetation production and the energy exchange rate. As such, it is an important input parameter for many physiological processes, ecosystem analyses and ecological modelling studies. Numerous direct and indirect approaches for estimating LAI have been developed (Jonckheere et al. 2004). Direct techniques include stratified-clip, leaf litter collection and dispersed individual plant methods. These direct measurement methods, through harvesting of vegetation, are destructive, extremely time- consuming and therefore difficult to use in large-scale study areas.

LAI is commonly estimated indirectly by measuring light penetration and vertical gap fraction by hemispherical photography (Jonckheere et al. 2004; Macfarlane et al. 2007; Osmond 2010), inclined point quadrat (Wilson 1963), allometric techniques (Gower and Norman 1991; Smith 127 et al. 1991), Multiband Vegetation Imagery (MVI) (Kucharik et al. 1997), LAI-2000 Plant Canopy Analyzer (Nackaerts et al. 2000), Tracing Radiation and Architecture of Canopies (TRAC) (Chen et al. 1997), SUNSCAN Canopy Analysis System (Vojtech et al. 2007), AccuPAR (Garrigues et al. 2008), and DEMON (Lang 1986). In addition to these indirect techniques, remote sensing shows the most accurate results on a larger spatial scale. However, this method requires technical experience with remote sensing software.

In this study, Leaf Area (LA) is calculated by using regression equations from easily measured tree traits such as diameter at breast height (DBH), tree height, and crown parameters, and then entering them into the i-Tree Eco application (UFORE model) along with additional data, e.g. CLE value. The UFORE model estimates maximum leaf area empirically derived from regression equations and the shading coefficient for deciduous urban species based on input variables from field data, and then applies correction factors to extend it to other tree types (Nowak 1996; Nowak et al. 2008). If the shading coefficient used in the regression equation does not exist for an individual species, genus average or 0.83 (the average shading coefficient for deciduous trees) is applied. However, a number of scholars use 0.88 for the average shading coefficient. LAI is applied to the ground area (m2) occupied by the tree to calculate Leaf Area (m2). For conifer trees (with the exception of pine trees), the average LAI per height-to-width ratio class for deciduous trees with a shading coefficient of 0.91 is applied to the tree’s ground area to calculate Leaf Area. According to Nowak (1996), conifer trees typically have about 1.5 times more LA than deciduous trees. For pine trees, the average shading coefficient (0.83) is used to calculate pine leaf area because, compared to other conifers, it has lower LAI and is similar to deciduous trees.

Regression equations for predicting leaf area based on shading coefficient and DBH:

lnY b  b X  b S Eq. 6. 12 0 1 2 Regression equations based on shading coefficient, dbh and crown height: lnY b  b H  b D  b S  b C Eq. 6. 13 0 1 2 3 4

Where, Y is leaf area (m2), X is dbh (cm), H is crown height (m), D is average crown diameter (m), S is the average shading coefficient for an individual species, and C is the outer surface area of the tree crown, can be calculated as below:

CDHD ( ) / 2 Eq. 6. 14

Leaf area then needs to be modified downwards depending on the tree condition (e.g. crown leaf dieback) where the range is between 0 (dead tree or 100 percent dieback) and 1 (excellent and healthy tree condition or no dieback). Trees are assigned one of seven condition classes: excellent (less than 1% dieback); good (1% to 10% dieback); fair (11% to 25% dieback); poor (26% to 50% dieback); critical (51% to 75% dieback); dying (76% to 99% dieback) and; dead 128

(100% dieback). The middle range, for example 11-25% dieback (fair condition), is given a rating of 0.82 or 82 percent healthy crown. Tree leaf area is multiplied by the tree condition factor to produce the final LA estimate (Nowak et al. 2008).

To estimate the shading effect of trees considering LA as its corresponding parameter, additional adjustments for open-grown or overlapping crown are required. There are a number of formulas that calculate crown competition factor. The iTree application modifies LA by using Crown Light Exposure (CLE) input data. As such, the CLE value must be collected during data collection to adjust LA per tree species. For trees with CLE=4-5, leaf area was based on equations for open-grown trees; for trees with CLE=0-1, leaf area was based on the LAI of closed canopy multiplied by canopy projected crown area:

Eq. 6. 15 ln(IX s ) 2 LA r k

Where, Xs is the average shading coefficient of the species and r is the crown radius. For CLE = 2-3, LA is calculated as the average leaf area from the open-grown (CLE = 4-5) and closed canopy equations (CLE = 0-1).

Normalisation: There is a correlation between LA and shade. As such, increasing the percentage of LA in each grid will increase the shade effect. Table 6.13 gives the normalisation value of the shading effect. Based on the results from calculations of the UFORE model, LAI and LA have been estimated for the individual tree in each grid. Then the shade effect ratio was calculated by rescaling the normalisation method known as z-score (Nardo et al. 2005) then Percentile calculator (1-sided known as NORMSDIST formula in Excel), was used to convert the z-score to a percentage. In this calculation, the highest record for LA consider as 100%, as shown below:

X LA   Eq. 6. 16 z_ score   

SE NORMSDIST( z _ score )*100 Eq. 6. 17

Where, SE is shading effect, XLA is the sum of LA of all individual trees in each grid, µ is arithmetic mean of LA, σ is standard deviation. The Normsdist procedure normalises the value of LA in each grid so that they all have an identical range (0% - 100%). According to Table 6.13, a high value of the LA ratio will achieve a higher baseline value.

Table 6. 13 Normalisation of the shading effect of trees Shading effect (%) Baseline value >80 5 60-80 4 40-60 3

129

20-40 2 <20 1 Improvement strategy/Future scenario testing: To achieve the highest possible leaf area per grid (100 m by 100 m), available growing space is filled preferentially with large trees first, then medium-sized trees, and finally small trees. Again, this planting strategy may not be practical or desirable in all situations, but the goal of this indicator is to establish a baseline for the maximum leaf area in each grid cell that could potentially provide higher performance in terms of shade among other attributes for this particular indicator.

Policies surrounding new urban tree selection should focus on trees that are well adapted to urban conditions while at the same time have the potential to provide the greatest shading benefit in summer. Where possible, and especially in sparsely treed urban locations, planting strategies should focus on providing shade during peak solar access periods. Under such conditions, preference for shading of west-facing structures will have the greatest magnitude benefit on mitigating rise in microclimatic temperature while reducing solar gain on buildings. For greatest impact, trees should be located 5m to 10 m from the building wall.

In instances where there may not be suitable soil conditions or space to sustain a shade tree, or where nearby obstructions create potential conflicts, vertical greening systems such as living walls or green façades (e.g. perennial vines, climbing plants) can provide a viable alternative and effective solution at mitigating building surface temperature resulting in lowering of indoor energy consumption. Cheng et al. (2010) in a field study in Spain concluded that green façades can reduce the interior surface temperatures at least 2°C, although this depends on various factors e.g. amount, density and species types (deciduous or evergreen) and façade orientation, supporting structures and level of maintenance.

INDICATOR C2: Air pollutant removal (Air quality improvement)

Description: In urban areas, poor air quality is of particular concern because it is a major cause of premature mortality due to respiratory and cardiovascular problems (Brunekreef and Holgate 2002; Rückerl et al. 2011). A dense urban area with a high canyon aspect ratio, which usually blocks wind flow, also traps pollutants between buildings. However, urban GI can directly and indirectly impact on air quality in four main ways (Nowak 2002; Nowak et al. 2006, p1). - Temperature reduction and other microclimatic effects - Removal of air pollutants - Emission of Biogenic Volatile Organic Compounds (BVOC)s and tree maintenance emissions - Reduction of building energy use

The removal of pollutants through either the absorbance of gaseous pollutants via leaf stomata (SO2, NO2, CO) and/or interception of particulate matter on leaves (PM10, PM2.5) directly improves air quality, as does decreasing outdoor air temperatures by shading and

130 evapotranspiration, which lowers ozone levels (O3) indirectly. The emission of many pollutants and/or ozone-forming particles can be influenced by air temperature. Although Biogenic Volatile Organic Compounds (BVOC) are emitted by tree canopies and can contribute to ozone formation and carbon monoxide by combining with NOx (nitrogen oxides) in the presence of sunlight, BVOC emissions are temperature dependent and trees generally lower air temperatures. Therefore, increasing tree cover by selecting species wisely through minimising the proportion of BVOC-emitters can lower overall BVOC emissions and thereby decrease ozone levels in urban areas (Coutts 2016).

Indirectly, air quality improvement occurs when air-conditioning use and related energy consumption for cooling and heating in buildings is reduced (through shading of buildings and wind modification), leading to lower air pollutant emissions from power plants (known as ‘avoided emissions’). A simulation of the meteorology and air quality in Los Angles by Akbari et al. (1996) found that daytime temperature reductions by 6°F (-14.45 C) with cool surfaces and trees would decrease air-conditioning usage, resulting in an estimated 10% reduction in smog levels. A similar study conducted by Taha et al. (2000) in Sacramento concluded that increasing tree cover and adding cool roofing lowered air temperature by about 3°F (-16.11 C) and reduced smog levels By 5.5% .

Calculation: iTree Eco Software is utilised to calculate air pollution removal by trees and 2 shrubs. iTree Eco v6 (2016) calculates six pollutants (CO, NO2, O3, PM2.5 and SO2; in gr/m per year as well as BVOC emissions). Particulate matter of less than 10 microns (PM10) is another significant air pollutant. Given that the latest version of iTree Eco analyzes particulate matter of less than 2.5 microns (PM2.5), PM10 has not been included in this analysis. PM2.5, because of the miniscule size of the particles, is generally more relevant than PM10 in discussions concerning air pollution effects on human health and the increased risk of lung cancer and cardio-pulmonary mortality (Coutts 2016). iTree utilises the Urban Forest Effects (UFORE) model to calculate the hourly dry deposition (i.e. pollution removal during non-precipitation periods) of CO, NO2, O3 and SO2 as well as the daily deposition of particulate matter of less than 2.5 microns. This calculation requires field data collection, hourly air pollution concentration data, and meteorological data. The amount of a given pollutant removed by urban green cover is calculated using the following equations:

QFST   Eq. 6. 18

Where, Q is the amount of a given pollutant removed by urban greenery (kg) during a given time period T, F is the pollutant flux (g m-2 s-1) and S is the total green cover in that area (m2). -2 -1 The pollutant flux (F, in g m s ) is calculated as the product of the deposition velocity (Vd; in m s-1) and the pollutant concentration (C; in g m -3) (Nowak 1994a):

131

FVC Eq. 6. 19 d

Deposition velocity (Vd) is calculated as the inverse of the sum of the aerodynamic (Ra), quasi- laminar boundary layer (Rb) and canopy (Rc) resistances.

1 V  Eq. 6. 20 d ()RRRa b c

Hourly meteorological data from the closest weather station are used to calculate Ra and Rb using standard resistance formulas (Killus et al. 1984). The algorithms for calculating Ra and Rb are given by Nowak and Crane (2000). Rc is related to leaf phenology and leaf area. Three variables are required to calculate Rc:

1R 1/ ( r  r )  1/ r Eq. 6. 21 c s m t

Where, rs is leaf stomatal resistance, rm is leaf mesophyll resistance, and rt is leaf cuticle resistance (Yang et al. 2005; Yang et al. 2008).

The UFORE model uses tree and shrub leaf area (LA) and determines the leaf area percentage for coniferous and deciduous trees by considering local leaf-on and leaf-off dates to account for seasonal leaf area variation. Transpiration and related pollution deposition of deciduous trees is limited only to the in-leaf period (Nowak and Crane 2000). Therefore, Nowak (1994a) classified canopy resistance (Rc) into leaf-on season daytime, leaf-on season nighttime, and leaf-off season, under the assumption that there is a distribution of 10% coniferous and 90% deciduous leaf surface area (LA). iTree Eco reports monthly air quality improvement in grams for individual pollutants that have been removed by urban greenery per year, where pollutant uptake into the atmosphere is equal to a given pollutant concentration (usually in g m-3) multiplied by boundary layer height (m) and then multiplied by the area of the site (m2). According to Nowak et al. (2006), minimum boundary layer heights were set to 150m during the night and 250m during the day.

The overall reduction of the five major air contaminants considered in this study is: 5 Eq. 6. 22 X AP AP i i1

Where XAP indicates the total amount of air pollution removal (gr per year); and APi represents the individual pollutants removed by GI in each grid (gr per year) (including five pollutants (i) CO, NO2, O3, SO2 and PM 2.5).

132

Normalisation: Increasing the percentage of green cover does not necessarily mean that the maximum amount of pollution will be absorbed. Nowak et al. (2006) suggests that, where canopy cover increases to 100%, trees remove less than 2% of total air pollution in urban environments and roughly 10% in more rural locations. Increasing canopy cover sometimes results in unwanted consequences in terms of air pollution uptake, e.g. increasing tree canopies alongside a deep canyon can limit wind flow, which causes higher pollutant concentrations as well as increasing the BOV emissions by trees and consequently raising ozone levels. Therefore, in addition to other factors, considering the effect of tree spacing in street canyons is critical, and larger tree spacing ensures better natural ventilation (Wania et al. 2012), consequently decreasing air pollutant concentrations (Vos et al. 2013).

Based on the study conducted by Nowak et al. (2006) across 55 cities, the urban green cover pollution removal values per unit canopy cover in an urban area range between 6.2 gr/m2 and 23.1gr/ m2 with the median pollution removal value per unit green coverage being 10.8 g/ m2.

However, in the case study of Parramatta CBD , the value of air pollution removed through GI ranges between 2.25 and 11.37 gr per m2 green cover with the median pollution removal value of 4.1 g/ m2. This value is out of the range reported by Nowak et al. (2006) . Therefore, for this specific study area, the baseline value reported by Nowak et al. (2006) that ranged between 6.2 gr/m2 and 23.1gr/m2 is not applicable. Therefore, the range between 6.2 gr/m3 and 23.1 gr/m2 can not be considered as baseline values. Hence, the normalisation procedure is applied and a standard range between 0 and 100 percent is assigned.

Table 6.14 presents the normalisation value of the air pollutant removal ratio. Based on the results from calculations of the UFORE model, the sum of all five pollutants removed by urban tree canopies has been estimated for the each tree in each grid. Then the air pollutant ratio was calculated by rescaling the normalisation method known as z-score (Nardo et al. 2005) then Percentile calculator (NORMSDIST formulas), was used to convert z-score to the percentage, as shown below:

X AP   Eq. 6. 23 z_ score   

APR NORMSDIST( z _ score )*100 Eq. 6. 24

Where, APR is the percentage of air pollutants removal, XAP is total air pollutants removed by individual trees in each grid, µ is arithmetic mean of total AP, σ is the standard deviation of total AP. The Normsdist procedure normalizes the value of AP in each grid so that they all have an identical range (0%- 100%). According to Table 6.14, a high APR ratio achieves a higher baseline value.

133

Table 6. 14 Normalisation of air pollutant removal ratio

Air pollutant removal (%) Baseline value >80 5 60-80 4 40-60 3 20-40 2 <20 1

Improvement strategy/Future scenario testing: Green infrastructure has the potential to contribute to the reduction of air pollutant concentrations in urban areas. While it is not always easy to plant trees, or other types of GI, in a densely populated urban area, selecting the right species and planting in the right location in available land is essential. Yang et al. (2008) show that the air pollutant removal rate of trees is double the rate for grasses, and compared with green roofs, a medium-sized tree can remove air pollution the equivalent of 19m2 of extensive green roof in one year (Currie and Bass 2008). This indicates that trees and secondly shrubs are more effective in capturing pollutants than grasses and groundcover species.

In addition, coniferous trees are more efficient in trapping pollutant particles than deciduous trees because they keep their leaves throughout the year and have higher total leaf surface areas to remove pollutants (Currie and Bass 2008). It is more advantageous to plant tree species with hairy or rough leaf surfaces, and to plant them closer to the pollution source that is, in most cases, next to busy roads.

Thus, for future scenario testing and proposed GI planning, the following assumptions are highlighted to increase the capacity of available land to uptake air pollutants:

- In available vacant land area, planting tree canopy is must be a higher priority than planting shrubs and grass to achieve the maximum air pollutant uptake benefits (Currie and Bass 2008) - however, this recommendation may be different for site specific conditions that requires an arborist’s advice; - Increase the number of large and healthy trees, especially coniferous and evergreen trees, e.g. native Pine species (for example hoop pines, Callitris, Wollemi pine) should be used more in urban areas, as they are most effective in capturing particles according to Sæbø et al. (2012) and Beckett et al. (2000); - Maintain existing native trees to support their health; - Use low-maintenance trees (reduces pollutant emissions for maintenance activities) (Nowak 2002); - Maximize use of low BVOC emitting trees (reduces ozone and carbon monoxide formation) species such as Jacaranda, Flindersia, Malus and Ginkgo; - Plant trees near buildings to conserve energy;

134

- Plant trees to shade parking lots (reduces vehicular VOC emissions; the cooler the car the lower the rate of petrol VOC evaporative emission rate from vehicles) (Nowak 2002); - Plant trees in polluted areas or heavily populated areas (maximizes tree air quality benefits); - Avoid planting species which are sensitive to pollutants (ensures tree health) (Nowak 2002).

INDICATOR C3: Carbon offset (Carbon storage and CO2 sequestration)

Description: According to Nowak and Crane (2001), trees and other types of vegetation, which act as a sink for CO2, are proven to be a particularly effective solution for decreasing the levels of atmospheric CO2 among other greenhouse gases. They can actively reduce CO2 in two ways: directly through the sequestration process and storage as a form of carbon within woody biomass and indirectly wind modification; and shading effects of trees around buildings can reduce the demand for heating and air conditioning (save energy), thereby reducing emissions associated with production of electric power (Dwyer et al. 1992; McPherson 1994). The Kyoto Protocol also recognises trees as a carbon sink and a valid means to offset greenhouse gas emissions to meet internationally-agreed emissions targets (Grace and Basso 2012).

During photosynthesis, atmospheric CO2 enters the leaf through surface pores, combines with water, and is converted into cellulose, sugars, and other materials in a chemical reaction catalysed by sunlight. Most of these materials become fixed as wood, although some are respired back to CO2 or used to make leaves that are eventually shed by the tree. The combined effect of photosynthesis and respiration results in net storage of CO2 by the tree (Nowak and Crane 2001).

‘Carbon dioxide sequestration’ refers to the annual rate of capture and storage of CO2 in plant over the course of one growing season. Sequestration depends on tree growth and mortality rates, which in turn depends on species composition and age structure of the urban forest (McPherson 1998). The term ‘carbon dioxide storage’ refers to the accumulation of woody biomass as trees grow over time. The amount of carbon stored in biomass is influenced by tree density and management practices (McPherson 1994).

Calculation: Carbon storage and sequestration in urban ecosystems is becoming a commonly used approach for managing climate change, although biomass allometric equations in the urban context still are not well documented (Aguaron and McPherson 2012; Grace and Basso 2012). Recently, a number of scholars have developed evaluation methods of carbon storage and sequestration in urban areas based on empirical field measured forest-based studies (Beets et al. 2012; Escobedo et al. 2010; Jenkins et al. 2003; McHale et al. 2009; Nowak et al. 2013; Schwendenmann and Mitchell 2014; Zhao et al. 2010). Carbon densities will tend to increase

135 with tree density and/or increased proportion of large diameter trees. There is evidence that urban forests are estimated to store nearly 50% less carbon than natural forests per hectare basis due to the lower density and younger age of trees in urban areas (McPherson 2010; Nowak and Crane 2002; Zhao et al. 2010). It indicates, two dominant factors that affect carbon storage density (tC/ha) are tree density (trees/ha) and diameter distribution. However, a study by Nowak and Crane (2002) found rates of carbon sequestration decrease as a tree matures if decline in health.

The general form of this allometric biomass equation consists of two variables: diameter at breast height; and tree height as follows:

Biomass()()()() DW or FW a  dbhbc  height Eq. 6. 25

Where, DW is dry-weight, FW is fresh-weight and dbh is diameter at breast height (conventionally in the US and Australia at 1.40 m). a, b and c are regression coefficients which are estimated from field studies for specific species.

Most equations produce dry-weight (DW) biomass, some equations compute fresh-weight (FW) biomass and are multiplied by species- or genus-specific conversion factors to convert to DW biomass. When a formula is not available for a given species the average of results from equations of the same genus is used. If no genus equations are found, the general equation for broadleaf or conifer is used to convert to DW biomass, multiply the broadleaf FW by 0.56 and the conifer FW by 0.48 (Center for Urban Forest Research (CUFR) 2008; McHale et al. 2009).

General broadleaf equation to calculate biomass (dbh is in cm): 2.310647 Eq. 6. 26 BroadleafBiomass() FW 0.280258 ( dbh )

General conifer equation to calculate biomass (dbh is in cm): 2.580671 Eq. 6. 27 ConiferBiomass() FW 0.05654 ( dbh )

These two equations estimate above ground biomass. So, the biomass stored below ground is added to above ground biomass multiplied by 1.28 (total biomass = 1.28 * above ground biomass). Based on the field research studies 50% of total DW biomass consists of carbon, so, to calculate stored carbon, total DW is multiplied by the constant 0.50. The sequestered CO2 rate in kilograms for a given species is then calculated as follows:

CO2 Biomass ( DW )  1.28  0.50  3.67  0.80 Eq. 6. 28

In this equation, 3.67 is the molecular weight of CO2 for converting stored carbon to carbon dioxide and 0.80 is the conversion factor between forest and open-grown urban trees. Stored 136

CO2 is usually reported in metric tonnes. Therefore, results from this equation are in kilograms and must be multiplied by 0.001 to convert to metric tonnes.

Limited sampling biomass studies have been conducted on urban species to verify the accuracy of estimates from these equations across a range of climates, growing conditions and tree structures distinguishing between open-grown and non open-grown trees (Aguaron and McPherson 2012; McHale et al. 2009; Pillsbury et al. 1998). Therefore, there is great uncertainty associated with the application of biomass equations across trees in the urban context. The CUFR Tree Carbon Calculator (CTCC) is a free Excel spreadsheet that was developed by US Forest Service researchers in 2008. This tool estimates carbon storage and sequestration for a single tree as well as annual heating and cooling energy savings by using climate zones reference cities in the U.S.A. and species information (McPherson et al. 2008; McPherson et al. 2016).

The CTCC developed the urban-based biomass allometric equations based on 26 species for trees growing in open urban conditions in California and Colorado. It is the only tool approved by the Urban Forest Project Protocol for quantifying CO2 sequestration. The Urban General Equations (UGEs) is another approach for calculating CO2 storage for broadleaf, conifer, and palm trees in urban areas by utilising existing imagery obtained by remote sensing. Aguaron and McPherson (2012) reviewed four methods to calculate carbon storage and sequestration including CTCC, UGEs, iTree Eco and iTree Streets. Their research indicated that iTree Eco, in comparison with the three other tools, produced lower results for carbon storage and sequestration estimation, maybe because iTree applied three subtraction factors including: 0.80 correction factor (to convert forest-based equations to urban condition), and also adjusted the result based on projected mortality rates, location and health conditions of canopies. The carbon storage estimates produced by i-Tree Streets and CTCC were the highest, while UGEs produced relatively low estimates of carbon storage.

In this study iTree Eco-v6 software (formerly known as the UFORE model) was utilised to calculate carbon storage and sequestration. The UFORE carbon model was developed from forest-based biomass allometric equations and tree growth modelling that relates tree dimensions (dbh and height) to tree volume or biomass. A 0.80 correction factor is used to calculate carbon storage and sequestration in urban area with open-grown trees (Nowak 1994b; Nowak and Crane 2002).

In iTree Eco, Hahn (1984) equations are applied to calculate biomass for deciduous trees greater than 94 cm dbh and coniferous trees greater than 122 cm dbh (Nowak and Crane 2002). Also, Eco considers standardised tree growth based on the number of frost free days and adjusts this base value based on tree condition, location and crown light exposure (CLE) to calculate sequestration (Nowak 1994b; Nowak et al. 2008). Figure 6.4 shows the data that needs to be collected for running the iTree Eco model to calculate carbon storage and sequestration (Nowak and Crane 2002; Nowak et al. 2008):

137

Main input data: DBH = diameter at breast height (1.4 m), Ht = total height, CLE = crown light exposure, D = crown dieback. Additional input data: Cw =crown width, Cb = crown base height, PCM = percent canopy missing.

Figure 6. 4 Schematic figure of required iTree input data

In the iTree software, growth rate is adjusted based on default values for tree health condition as follows: fair to excellent condition – multiplied by 1 (no adjustment); poor condition – 0.76; critical condition – 0.42; dying – 0.15; and dead – 0. These growth adjustment factors are based on percent crown dieback and the assumption that less than 25% crown dieback had a limited effect on dbh growth rates (Nowak et al. 2008). Crown light exposure (CLE) provides information on the number of sides of the tree receiving sunlight and ranges from 0 (no full light) to 5 (full light from top and 4 sides). Finally, gross sequestration is estimated from annual tree growth. Net sequestration incorporates CO2 emissions due to decomposition after tree death. Emissions are based on the probability of the tree dying within the next year and being removed.

Normalisation: Urban forestry strategies offer potential contribution of green cover to mitigate carbon dioxide and climate change. It is evident that a single healthy mature tree in an urban area can sequester carbon dioxide at a rate of 21.77 kg/ per year (City of Sydney 2013). Therefore, the carbon sequestration ratio is calculated by dividing the amount of carbon

138 sequestered (kg/yr) per grid by potential carbon sequestration value per grid cell which is obtained by multiplying the maximum number of trees per hectare in the study area by 21.77:

(,)ik C Eq. 6. 29 i/ Grid CSR 100 21.77 Nmax/Grid

Where, CSR is carbon sequestration ratio, C(,)ik is carbon sequestration value for individual tree in each grid that was calculated by iTree Eco platform and Nmax/Grid is the maximum number of trees per grid (assume a 10-metre spacing between trees).

This method is similar to the proposed method by McPherson et al. (2013) to identify and compare capacity of carbon sink in each grid. They suggested using percentage of green cover as a maximum capacity of carbon sink in each plot. In the case study of Parramatta, the average value for CO2 sequestration was 17.51 kg/yr - lower than this proposed value of 21.77 kg/ per year for mature trees. Therefore, the value in the formulas was adjusted to 17.51 kg/yr for the calculations in this thesis.

Baseline values are given in Table 6.15.

Table 6. 15 Normalisation of carbon sequestration ratio

Carbon sequestration ratio (%) Baseline value >80 5 60-80 4 40-60 3 20-40 2 <20 1

Improvement strategy/Future scenario testing: One of the keys for maximising CO2 sequestration is the selection of species well adapted to the climate. The growth rate, life expectancy and maintenance required must also be taken into account for species selection. In addition, the other main factor for increasing carbon storage and gross sequestration per hectare is increasing tree density while selecting species with higher rate of sequestration. In general, variation of carbon storage and sequestration between conifer and deciduous trees indicates that evergreen trees continue sequestering and storing carbon for the whole year while deciduous trees drop their leaves annually and sequester carbon in half of the year while storing carbon during leaf-off seasons in their wood biomass. Younger trees also have the higher annual rate of sequestration during the growing season while large trees with a healthy leaf area provide the best storage of carbon in their trunks.

INDICATOR C4: Reduced building energy used for cooling and heating

Description: Trees can help to reduce energy consumption in two ways: 1) shade effect; and 2) climate effects by way of blocking or directing wind (McPherson and Simpson 1999).

139

However, each parameter is affected differently based on tree type, size and planting location. Trees provide windbreaks, which can lower wind speeds resulting in reduced loss of heat in the winter. Planting in the ideal location and with the appropriate distance from the building is important to energy savings measured by this indicator. For example, the CUFR Tree Carbon Calculator (CTCC) and iTree Eco only consider tree height of >=6.2 m and a distance of up to 18.3m from one- or two-storey buildings to be relevant when calculating energy savings. Any tree that is smaller than 6.2m in height, or is further than 18.3m from a building, is considered to have no effect on building energy use.

Calculation: iTree Eco application uses a method that was developed by McPherson and Simpson (1999) to report the amount of carbon avoided from power plants due to the presence of trees with regards to two categories: heating (in MBTU); and cooling (Kwh), including both positive and negative effects. Various data must be recorded for application input including: tree size; condition and type (deciduous or evergreen); and distance and direction (azimuth) from buildings (the latest version of iTree is able to calculate the energy effects of any given tree on the three closest buildings to it). Information about building type, materials and residents' energy usage (heating and cooling) have been set as a default based on data from the United States. This building type information has not yet been adjusted for the Australian version.

Therefore, this model can run based on the Australian species and climate but results for energy saving seems to be lower than expected due to a difference in building types and materials. By considering all these limitations, to calculate heating and cooling benefits of urban tree canopy, iTree Eco was utilised. Analysis of the results indicates that heating value due to tree windbreak effects was too insignificant to compare with cooling effect. Therefore, saving energy for heating was not considered in the GIS modeling. The iTree procedure also only considers a tree height of >=6.2 and a distance of up to 18.3m and the impact of trees up to second floor of building.

Normalisation: Based on the results from calculations of the UFORE model, the sum of cooling benefits of the individual tree (Kwh) have been estimated in each grid. Results are normalised by calculating the z-score in each grid (Nardo et al. 2005) and the Percentile calculator (NORMSDIST formulas) was applied to convert the z-score to a percentage, as shown below:

X cool   Eq. 6. 30 ESC  NORMSDIST *100 

Where, ESC is energy saving for cooling ratio, Xcool is total cooling benefits (kWh) provided by individual trees in each grid, µ is the arithmetic mean of total cooling (kWh) value, σ is standard deviation of total cooling value. The Normsdist procedure normalises the value of Xcool in each grid so that they all have an identical range (0%-100%). Table 6.16 presents the normalisation value of energy saving regarding cooling due to the shading effect of trees.

140

Table 6. 16 Normalization values for energy use reduction - only cooling (mature tree height)

Energy saving - cooling (%) Baseline value >80 5 60-80 4 40-60 3 20-40 2 <20 1 Improvement strategy/Future scenario testing: As a rule of thumb, in the southern hemisphere, planting healthy, mature trees closer to a building on its south, east and then west sides will provide the greatest impact in reducing energy use needed for cooling during the summertime. However, excess shade will increase energy use needed for heating during the winter. Therefore, it is recommended to plant native coniferous or evergreen tree species on the east and west sides and deciduous trees on the south and north sides of a building. In addition, planting evergreen trees is a good strategy to block wind in the winter and thereby reduce energy used for heating. However, as mentioned in the previous section, this effect is minimal compared to the effect trees have in regards to helping with cooling in the summer. Thus, the priority when planting trees in Australia is to assist in cooling.

INDICATOR C5: Hydrological regulation

Urbanisation alters the hydrological cycle of an area due to an increase in the percentage of impervious surfaces. This reduces soil infiltration rate as well as it increases water flow speed and runoff. GI, comprising both structural and non-structural stormwater Best Management Practices (BMPs), has been widely recognised as the best sustainable stormwater management system solution. This system helps to protect, restore and imitate processes of natural water cycle systems and it is generally accepted that Stormwater Management (SWM), or in Australia known as water sensitive urban design (WSUD), is one of an emerging number of climate change adaptation strategies (Lee and Heaney 2003).

Therefore, it is essential to integrate water cycle management into green infrastructure planning as an adaptive planning strategy for resilient and sustainable city planning. Green roofs, bioretention, rain garden water harvesting and other types of GI can be integrated as a part of a network to slow down water flow and improve water quality, as well as hold, mitigate, retain and infiltrate runoff volume. Two sub-indicators have been selected to assess the hydrological performance of green infrastructure: improving water quality; and reducing runoff by intercepting precipitation. The next two sub-sections describe their calculation.

INDICATOR C5-1: Water quality

Description: Land development is affecting water quality and the quantity of watersheds, mostly due to the conversion of pervious surfaces to sealed and impermeable surfaces. This increasing imperviousness proportionally increases polluted runoff. The common pollutants in storm water runoff are sediments, nutrients, heavy metals, oxygen-demanding material, pathogens, petroleum hydro-carbons and toxics. Removal of storm water pollutants can be 141 done through treatment processes including sedimentation, flotation, filtration, adsorption and biological degradation. Communities that are using Best Management Practices (BMPs) on a neighbourhood and municipal scale can reduce the possibility of storm water runoff, pollution and flooding.

Planting tree canopies along a street as a storm water treatment practice can increase pollutant uptake and reduce storm water runoff via rainfall interception and enhanced soil infiltration, soil stabilisation as well as with minimum effect through evapotranspiration (ET) (Pitman et al. 2015). In general, in comparison with grasses and shrubs, trees perform far superior in improving water quality because they can transform pollutants through their extensive root systems, as well as increase the infiltration rate by root penetration through typically impermeable urban soil layers into more permeable soil layers.

Calculation: For this study, the three most common water contaminants were selected to be included in the model: total nitrogen (TN); total phosphorus (TP); and total suspended solids (TSS). In accordance with site issues and project objectives, all three water contaminants, or only one of them, can be included in the assessment model. Which of these contaminants can be removed by GI types are considered based on the land-use, land-cover pollutant loading rate.

Calculating surface water quality is complicated and varies depending on a range of rainfall events and catchment characteristics in addition to typical factors such as pollutant loads, land- use and impervious area fractions. To calculate pollutant loads (LP), the formula - known as the ‘Simple Method’, originally proposed by Schueler (1987) - is used in this study. This method utilises two equations to calculate pollutant loads. First, the runoff coefficient (Rvu) for each land-cover type is required to calculate as follows:

Eq. 6. 31 RIvu0.05  (0.009  u )

Where Rvu is runoff coefficient for land use type u, and Iu is percent of impervious cover (for example, if an area is 75% impervious or sealed then Iu=75). Iu for each land use type was derived from the Model for Urban Stormwater Improvement Catchment (MUSIC) guideline (Water Melbourne 2016).

The runoff coefficient (Rvu) is a dimensionless factor relating the amount of runoff to the amount of precipitation received. It is a larger value for areas with low infiltration and high runoff (asphalt and concrete pavement, steep slope), and lower for pervious and well-vegetated areas (forest, flat land). Therefore, permeability, gradient, the soil type and land use are important factors to measure runoff coefficient.

Then, the pollutant load is calculated with the following equation (Schueler 1987):

142

6 Eq. 6. 32 LPPRCAP(  J  vu  u  u )  10 u

Where LP is the annual pollutant load (kg) in each grid cell, P is precipitation (mm/year), PJ is the ratio of storms producing runoff (default = 0.9), Rvu is the runoff coefficient for land use type u, Cu is Even Mean Concentration (EMC) of pollutant in urban runoff for each land/ land 2 use type cover (mg/litre), and Au is the area of each land use type u, (m ).

The EMC can be estimated from site specific or local, regional and national data, or obtained from Table 6.17. These default EMC values presented in Table 6.17 are based on an extensive literature review in Australia and worldwide by Fletcher et al. (2004) on stormwater quality monitoring studies. Fletcher et al. (2004) recommend the range of values for EMC by giving the priority to the findings of local studies (those reported data from Sydney or NSW) that has derived findings from the greatest amount of water quality sampling data (Fletcher et al. 2004).

Table 6.17 also demonstrates the percentage of impervious cover (Iu) for individual land use/land cover recommended by Water Melbourne (2016) to include in runoff coefficient (Rvu) equation.

1 Table 6. 17 Event Mean Concentration (EMC) (Fletcher et al. 2004) and runoff coefficient (Rvu) (Water Melbourne 2016)2 in individual land uses Recommended typical value – Event Mean Concentration (EMC)1 Percent of Land use/Land cover TSS impervious Total Phosphorus Total Nitrogen 2 (mg/l) cover (Iu) % (mg/l) (mg/l)

Road 0.5 2.2 270 60-90 Roofs 0.13 2 20 75-95 General urban 0.25 2 140 60-90 Residential 0.25 2 140 50-80 Industrial 0.25 2 140 70-95 Commercial 0.25 2 140 70-95 Mixed urban/rural 0.25 2 100 40-70 Rural 0.22 2 90 10-30 Agricultural 0.6 3 140 5-20 Forest/Natural 0.08 0.9 40 0-5

After calculating the runoff coefficients and pollutant loads, the water quality improvements by implementing series of GI types can be calculated, as follows (Blick et al. 2004; Edwards and Miller 2001):

WQI L (1  R )  R  % A Eq. 6. 33  p GI(n 1) GI ( n ) GI ( n ) n

Where, WQI indicates the annual water quality improvement ratio; LP is land-use pollutant loads in each grid (mg per year); and RGI is the relative percentage of pollutant reduction by GI type n; and AGI is the area of GI type n that works as BMP solution to improve water quality.

143

Table 6.18 shows the percentage of pollutant reduction (RGI) by tree canopies if planted over impervious cover (e.g. footpath planter, roadway, and boulevard) or over grass. For example, a tree planted on grass will reduce a TN load 23.8%, whereas the reduction would be 8.5% if the tree were planted over an impervious surface. Information about other types of GI that help to uptake water pollutants effectively (e.g. vegetated bioretention system, swale, permeable paving etc) are explained later in Table 6.21.

Table 6. 18 Tree canopy relative land use loading rates based on underlying land-use land-cover (Cappiella et al. 2016b; Hynicka and Divers 2016). Tree canopies (RGI) Total Nitrogen removal Total Phosphorus removal Total Suspended solids rate (%) rate (%) removal rate (%) Canopy over grass 23.8 23.8 5.8 Canopy over impervious 8.5 11.0 7.0

Normalisation: In order to minimise stormwater pollution from the site receiving runoff, Water Sensitive Urban Design (WSUD) principles should be considered into the stormwater infrastructure treatments and landscape design phase of future development. Based on the data from Tables 6.18 and 6.21, the maximum percentage of water pollutants removed by each GI type is 80%. In addition, the NSW Government has set a target for current Best Practice for stormwater quality at 80%, 65% and 45% reduction for annual loads of TSS, TP and TN, respectively (City of Sydney 2012). This target is quite similar in Queensland and in South Australia (80%, 60% and 45%), while Victoria maintains a slightly lower phosphorus target (80%, 45% and 45%). However, the national target in Australia is 80%, 60% and 45% reduction for annual loads of TSS, TP and TN, respectively, with the 90% reduction in annual loads of gross pollutants/litter target value (Myers et al. 2011).

Table 6.19 indicates the comparison of the quality of urban runoff in Australia to global figures.

Table 6. 19 Quality of urban runoff in Australia and global (Wong et al. 2000).

Pollutant Australian range of Global range Global range of Australian standard pollutant mg/l pollutant load (mg/l) concentration (mg/l) (kg/ha/yr) TSS 20-1000 50-800 70-1800 <10% change TP 0.12-1.6 0.1-3 0.4-3 <0.01-0.1 TN 0.6-8.6 2-6 3-10 <0.1-0.75

Among the three pollutants, TSS is the most significant one. It causes damage to aquatic life, creates water turbidity, and acts as vehicle to transport other pollutants. BMPs are frequently used to remove TSS, which can best be controlled by minimising erosion. To control TSS, BMPs should include slowing the water flow through detention ponds or vegetation. Infiltration can also be an effective approach. In the pilot study in Parramatta watershed, TSS as a major site issue is reported to be at the high level of risk during heavy rainfall and high windy conditions (Khan and Byrnes 2016), thus, TSS reduction would be the preference during water 144 quality improvement process. Therefore, identifying priorities based on the risk assessment and goals of the project are critical in selecting, designing and implementing the effective type of technical GI.

To set up a baseline values for water quality improvements, five reference levels were equally assigned in the range between 0 and 100 percent, as shown in Table 6.20.

Table 6. 20 Normalisation of water quality improvements ratio

Water Quality Improvements (%) Baseline value >80 5 60-80 4 40-60 3 20-40 2 <20 1

Note that this baseline value will apply when installing various BMPs on the site. For an existing urban GI, if the site does not have any technical GI or BMPs (Table 6.21), these baseline values are not applicable. For example, in the case study of Parramatta, street trees and green spaces are the only GI types: therefore, the existing GI types do not contribute to any effective removal of pollutants.

Improvement strategy/Future scenario testing: Recent trends in stormwater management have identified innovative practices and techniques that are showing promising improvements in the management of stormwater quality. These techniques include the adoption of best management practices (BMPs) by considering Water Sensitive Urban Design (WSUD), or the equivalent term in the United States known as Low Impact Development (LID) strategies. These address stormwater at its source, such as impervious surface reduction and utilisation of vegetated swales, biofilters and porous pavement. However, in designing BMP, the goals of the specific project are important to consider. For example, if phosphorous removal is a priority on the site, bioretention systems, such as rain gardens, will need to be designed intentionally - possibly to the exclusion of some of the soil amendments (e.g. compost) that have been known to also release phosphorous.

Table 6.21 describes various types of GI that contribute to nitrogen and phosphorus reduction. The most effective phosphorus removal BMP is bioretention and sand filter followed by the wet detention basin. The least effective phosphorus removal BMP is a dry extended detention basin. The highest nitrogen removal rate is produced by wetlands while the lowest is dry extended detention basins. It should be noted that nitrogen and phosphorus are required by plants and algae to grow. However, high levels of these nutrients lead to eutrophication, which leads to increased rates of photosynthesis and causes algae blooms. Natural waters have 0.02 mg/L phosphorus which reflects that it is a limiting factor for plant growth. Thus, to limit eutrophication, phosphate concentrations should be limited to maximum 0.1 mg/L (Behar et al. 1997). In addition to selecting an appropriate GI type, grass should be minimised in favour of other vegetated cover because a significant amount of nutrients are in stormwater runoff due 145 to fertilisation of lawns. In Australia, native plants tend to have low phosphorus demand, hence high phosphorus in soil is more likely to act as a pollutant.

Table 6.21 also indicates that the highest removal rate of TSS occurs through implementing bioretention, such as rain gardens. Constructed wetlands, vegetative filters, and infiltration filters are also found to have good removal rates for total suspended solids (Blick et al. 2004).

Table 6. 21 Typical TN, TP and TSS Removal rates, drainage size and requires space for different BMP types (Fletcher et al. 2004*; North Carolina Division of Water Quality 2007)

TN TP TSS Best Management Size of Runoff removal removal removal Removal Space Practice (BMP) drainage volume efficiency efficiency efficiency Ability required types area reduction (%) (%) (%) Bioretention (e.g. 35 45 85 High S High Possible rain garden) Stormwater 40 35 85 Med S-L High Yes wetlands Wet detention 25 40 85 Med M-L High No basin Sand filter 35 45 85 High S Med Possible Filter strip 20 35 25-40 Med S Med No Grassed swale 20 20 35 Low S Low No Restored riparian 30 35 60 Med S-M Med No buffer Infiltration devices 30 35 85 High S-M High Possible Dry extended detention 10 10 50 Med S-L Med Yes basin Permeable pavement 60-80* 40-80* 70-100* Low S-M Low Yes system Vegetated roof 0 or 0 or 85* Low S High Yes (green roof) negative negative

Typically, to achieve a gross removal target (90% reduction of annual load), or individual targets for specific pollutants in the site, a single BMP type will not be able to meet the target by itself. Therefore, a number of treatments (BMPs) will be required and the selection, configuration and order in which land use they are installed are important to consider. In such cases, it is recommended that the best arrangement of BMPs installation would be in ascending order of removal rate from upstream to downstream. For example, the BMP with the highest TSS removal rate would be more effective to be located at the downstream (Blick et al. 2004).

INDICATOR C5-2: Avoid surface runoff

Description: Surface runoff transfers pollution to streams, wetlands, rivers and oceans. GI reduces the proportion of runoff by combined effect of its ability to intercept, evapotranspire and infiltrate water into the soil. A typical tree canopy and associated uncompacted soil can hold up to 2.54 cm runoff from an impervious surface area which is typically greater area than 146 area under the tree canopy (Marritz 2011). Modelling studies demonstrate that, as green cover in a neighborhood or watershed increases, consequently runoff decreases (and the inverse assumption is also true). Runoff reduction studies at watershed scale often describe results in terms of percentage of annual runoff reduction due to percent of tree coverage in the watershed (to compare with the runoff generated if trees were not present).

These results can be translated into a percent runoff reduction per unit area of canopy if watershed areas are provided in the studies. However, not all studies are conducted on an annual basis and the results (streamflow measured at the watershed outlet) reflect not just the effect of trees on the watershed but also the cumulative effect of all other land-cover types and watershed features. In the literature, the runoff reduction attributed to urban trees varies based on various factors such as site condition and climatic factors. This variation ranges from 2.6 to 88.8% and each study has a unique approach to quantifying runoff reduction from individual plots or sites to watersheds. Runoff reduction attributed to natural forests is similarly wide- ranging, with values from 8 to 80% from studies of water yield before and after deforestation of forested catchments (Cappiella et al. 2016a, 14).

Calculation: To calculate runoff reduction by GI types, iTree Eco is utilised. iTree v6 can estimate six hydraulic variables provided by trees/shrubs including: potential evaporation; potential transpiration; evaporation; transpiration; rainfall interception; and avoided runoff. In this calculation, annual rainfall interception estimated based on the physics water balance model developed by Wang et al. (2008) and improved later by Hirabayashi (2013a). iTree Eco measures annual avoided runoff provided by single tree’s rainfall interception. In the water balance model (Hirabayashi 2013a) it is assumed the fraction of pervious to impervious cover is 74.5% to 25.5% (Nowak and Greenfield 2012). This model is complicated and this study does not aim to explain the equations behind the iTree Eco hydrology model. However, it worth mentioning that the main structure of the model is based on the water balance equation as below (Hirabayashi 2013b):

P In  S  R  E Eq. 6. 34

Where, P (m3) is precipitation, In (m3) indicates infiltration to the pervious surface, S (m3) is depression storage on both impervious and pervious surface (maximum storage value is defined 0.0015 (m) and 0.001 (m) respectively), R is runoff from impervious cover, and E is evaporation. To determine infiltration and ground depression storage, the area of the impervious (VG) and pervious (VA) cover is multiplied by the depth of the water to convert to volume. A proportion of gross precipitation (P) is intercepted by canopy leaves, branches, and trunk, and evaporates back into the atmosphere (E), this is known as rainfall interception (I). Then, annual avoided runoff volume (R avoided) can be calculated as (Hirabayashi 2013b):

Eq. 6. 35 Ravoided  I  In  S 147

Where, I (m3) is rainfall interception depth which converted to volume by multiplying by VA (m2). In and S are both in volume (m3).

Figure 6.5, adapted from Hirabayashi (2013b), clearly demonstrates input variables that are used in the iTree Eco model to calculate avoided runoff.

Figure 6. 5 iTree Eco- precipitation interception model diagram (Hirabayashi 2013b)

Normalisation: Markart et al. (2006), based on runoff coefficient value and roughness of land cover, classified surface runoff ratio in seven classes (e.g. 0-10% ranked as high infiltration potential and very low volume and (>75% is high runoff potential, and 100% indicates sealed or wet areas (Table 6.22).

Table 6. 22 surface runoff ratio

Surface runoff ratio Class (% of precipitation) 0 0 >0-10 1 10-30 2 30-50 3 50-75 4 75-99 5 100 6

148

The runoff reduction attributed to urban trees ranges from 2.6 to 88.8% (Cappiella et al. 2016a). Based on the study by Markart et al. (2008), baseline values for avoided surface runoff ratio are assigned as shown in Table 6.23.

Table 6. 23 Normalisation of avoided surface runoff ratio

Avoided surface runoff Baseline value % >90 5 70-90 4 50-70 3 25-50 2 <25 1

Improvement strategy/Future scenario testing: It has evident in WSUD best practice that planting large trees canopies over impervious surfaces, such as a parking lots or streets, has a considerable impact (up to 8 times impacts) on decreasing runoff because it reduces peak flows in urban settings. In addition, dense vegetation such as woodland intercepts and helps infiltrate rainfall, thereby reducing runoff volumes and rates. Highly porous or permeable soils also can rapidly infiltrate rainfall generally produce less runoff volume than soils with more restrictive infiltration. Table 6.24 presents runoff reduction performance for various BMPs. A range of values represents the median and 75th percentile runoff reduction rates based on the literature search by Hirschman et al. (2008). According to this table bioretention and infiltration devices have higher capability to reduce runoff whilst wet ponds, wet swales and stormwater wetland have an insignificant effect on runoff volumes (Battiata et al. 2010; Hirschman et al. 2008).

Table 6. 24 Runoff reduction capabilities for various BMPs (Battiata et al. 2010; Hirschman et al. 2008) Best Management Practice (BMP) types Runoff Reduction (%) Bioretention (e.g. rain garden) 40-80 Soil amendments 50-75 Extended detention pond 0-15 Infiltration devices 50-90 Wet swale 0 Dry swale 40-60 Grass channel 10-20 Permeable pavement 45-75 system Vegetated roof (green roof) 45-60 Rooftop disconnection 25-50 Raintanks and Cisterns 40

INDICATOR C6: Biodiversity protection and enhancement 149

Conserving urban biodiversity (plants, animals and habitats) has the potential to be a major benefit and function of GI by providing effective connectivity and networks of green spaces known to restore fragmented habits as well as by contributing to species diversity and health (Benedict and McMahon 2002; Benedict and McMahon 2012; Tzoulas et al. 2007). Hostetler et al. (2011) suggests that conserving and enhancing urban biodiversity through implementation of GI is only the first step, while the government and city planners should use a systematic approach to engage with communities and convince them that it is in their best interest to take steps to implement GI in their own spheres of control.

In this regard, two sub-indicators have been selected to assess the performance of green infrastructure in protecting and enhancing biodiversity in urban areas. Habitat connectivity (1) promotes the conservation of biodiversity (only fauna to calculate in this study). This indicator indicates structural connectivity of the green network. (2) Species diversity (only flora is calculated in this study). The next two sub-sections describe proxy indicators and calculation methods.

INDICATOR C6-1: Habitat connectivity (Habitat suitability)

Description: Landscape connectivity as a key component of green infrastructure planning has been defined as ‘the degree to which the landscape impedes or facilitates movement among resource patches’ (Taylor et al. 1993); or the maximum capability for species (wildlife and species) to move easily within a natural landscape (in this study). Generally, it is accepted that landscape connectivity plays an important role in dispersing organisms among habitat patches, conserving biodiversity (Tischendorf and Fahrig 2000) and ensuring that ecological processes continue (Leibold et al. 2004; Wright et al. 2003). Connectivity can be characterised as either structural or functional/biological. Structural connectivity describes the physical elements of the landscape that influence the distribution of species. It also explains spatial relationships among habitat patches such as inter-patch distances, size and configuration of a patch, and pattern and availability of a corridor. Functional connectivity measures the ability of organisms to move among patches based on the surrounding landscape and its spatial configuration (Taylor et al. 2006). It does this by considering the movement habits of an organism and the spatial and temporal characteristics of the landscape that support or limit movement.

The main strategy of GI planning is to increase habitat connectivity and mitigate the effects of land fragmentation due to anthropogenic pressures. One such strategy involves establishing and maintaining greenways and green corridors that link green open spaces together and provide increased habitat connectivity and gene flow, while simultaneously providing recreational opportunities. A large, interconnected network of greenways reduces habitat patch isolation and enables species migration and genetic movement between patches.

Calculation: All assessment models and metrics vary in complexity, from those that consider only the structural characteristics of the landscape, or functional connectivity to those that

150 assess landscape connectivity as a whole. Structural connectivity metrics relate to characteristics of habitat such as habitat suitability (distance, distribution, and patterning of habitats). Two types of distance metrics are: (1) nearest-neighbour distance metric (Moilanen and Nieminen 2002), which is more common and uses the Euclidean shortest path measurement between a focal habitat patch (or patches) and its nearest neighbourhoods, and (2) neighbourhood distance metric, which considers the distances between a focal habitat patch and all other patches. Another method is the spatial pattern metric, which consists of a set of indicators based on landscape characteristics that influence connectivity including shape index, patch perimeter, core area, patch area, fractal dimension, extent shape, and spatial arrangement of landscape elements (McGarigal 2002; McGarigal et al. 2002).

Structural connectivity is appealing in its ability to provide an efficient and straightforward assessment method. However, it should be used with caution because it does not consider species dispersal abilities and it might not be accurate in real-life scenarios (Jacobson and Peres-Neto 2010; Taylor et al. 2006). Potential connectivity is another metric that combines landscape attributes (patch and corridor size, shape, quality and distribution of patch) with the dispersal biology of focal organism(s) (Calabrese and Fagan 2004; Schumaker 1996). Graph theory (Urban and Keitt 2001), least-cost analysis and circuit theory models (McRae et al. 2008) are three key modelling techniques used to assess potential connectivity on a large spatial scale. On a smaller spatial scale, buffer radius and incidence function are two common techniques used to estimate potential connectivity.

In this study, the least-cost model has been adopted to calculate potential connectivity and effective distance between habitats (or ecological resource patches). To execute this method, the resistance surface or ecological cost of travel across habitats is required for the entire study area. The resistance surface (or cost-surface) is typically in a raster format that is created by combining multiple GIS datasets that represent the relative ease or difficulty associated with passing through any given pixel. It is sometimes known as a ‘friction surface’. In a resistance surface GIS layer, higher values indicate greater amounts of resistance (or friction) to species movement. In other words, it is expected to see less or zero species in this area with high resistance value. This will help to identify how, where, and to what extent certain patches are connected to one another or disconnected.

To calculate the resistance surface for fauna species, the scientific literature indicates that a number of main factors including land-use/land cover, elevation, slope, distance to water bodies and physical barriers such as roads and highways are important (Zeller et al. 2012). The GIS model for linkage design and habitat connectivity is typically built based on these factors. Even if they do not encompass a habitat’s characteristics and species requirements comprehensively, including all four of these four factors will increase validation of the model:

- Land-use/Land cover: These two are the most important habitat modelling factors for many species because they reflect food availability, potential shelters and other resources as well as the level of human disturbance on vegetation clearing. Depending 151

on data availability and literature, both factors can be included in the assessment metrics. - Physical geography information (e.g. elevation, slope, aspect): This factor is a determinant of land cover ranges and surface characteristics. Elevation is important to include in habitat modelling when literature states that specific species occur within a certain range of elevation. Then, the values should be classified into three or five ranges (below, within and above the elevation limits). Elevation is the basis for calculating several other derived variables such as slope, aspect and topographical position. Digital Elevation Model (DEM) data can be utilised to analyse elevation, slope and aspect through ArcGIS. These variables determine ease of movement of individual species. In the modelling procedure, slope and topography analysis are known as cost of movement. - Distance to water (rivers, streams, lakes, reservoirs etc): Usually, certain species colonise within a certain distance of water, which is correlated to their migratory movement, food and water requirements. But in some models, streams count as a physical barrier for movement. So, it is very subjective to include this factor in the model and it depends on individual species characteristics. - Barriers/distance to roads: When an organism is moving location, it usually encounters difficulties with physical barriers such as fences, roads or streams. Some habitat models classify these barriers and assign scores based on road types, density and distance from roads. This generates the resistance value that represents the probability of animals’ existence.

In addition to these factors, other variables can be taken into account, e.g. human population density, soil types, presence of food and cover, as well as type, size and shape of habitat patches that influence the distribution of species or colonies of the specific fauna species settled around specific plant species.

Before combining input layers (land-use/ landcover, elevation, slope, distance to water and roads etc.), values need to be standardised in the range of between 1 and 5 (Table 6.25). This range of scores was assigned to individual selected factors based on habitat suitability, which was derived from scientific literature, experts’ input and specifically based on occurrence species data in the study area of Parramatta. A score of zero should only be given to a factor when that factor is not applicable for specific fauna species, even if the other factors were optimal for that species. Table 6.25 represents a generic resistance score based on the literature review and a GIS data fauna occurrence survey between 2000 and 2014 in study area of Parramatta LGA.

In study area of Parramatta: - Scoring Landcover was based on surface classification in this study through image segmentation. And, as a rule of thumb, water and dense trees achieve a higher score as they have the potential to provide shelter and nests for species.

152

- According to the biodiversity strategy in the Parramatta Local Government Area (Parramatta City Council 2015, p.38), the six land-use planning zones, including Environmental conservation, Environmental management, Natural waterways, Public recreation, Private recreation, Recreational waterways, provide the greatest opportunity for protection and retention of biodiversity. Therefore, a score of 1 was assigned to these six land-use zones and the rest of the land-use zones were scored based on species occurrence data, accordingly. - The species occurrence value for elevation in the study area was between 0 and 65 metres. However, the grey-headed flying fox (Pteropus poliocephalus), which is one of the dominant species in Parramatta LGA, was found at an altitude lower than 200 m. Thus, the resistance score value was ranged between 0 and 200 metres. - Resistance score value of slope has been derived from Kushari et al. (2015), which can be seen with our slope analysis results through GIS . - Distance of fauna species occurrence to water bodies was recorded between 0 and 800 metres with the average value of 200m and the maximum number of species recorded to be present in 50 metres far away water bodies. Therefore, the resistance score was set in the 0-200 range, with levels dividing into five equal reference levels. - Distance to roads varies between 1.6 and 480 metres with the average value of 200 metres. Hence, five reference levels were equally assigned in the range of 0 and 200 m, as shown in Table 6.25.

Table 6. 25 Generic resistance scores of habitats suitability (based on recorded fauna species occurrence data in Parramatta LGA).

Factors Class i Resistance score (Si) Water 1 Dense tree canopy 2 Landcover Open spaces- scattered trees 3 Shrub/grass 4 Built surfaces (roads, buildings etc) 5 Environmental conservation, Environmental management, Natural waterway, Public recreation, Private recreation, Recreational 1 waterway Low density residential 2 Land-use Mixed use, Neighborhood centre, Special activities 3 Medium density residential 4 High density residential, Industrial, Infrastructure, Business 5 development 0-25 m 1 25-50 m 2 Elevation 50-100 m 3 100-200 m 4 >200 m 5 0-3% 1 3-9% 2 Slope 9-16% 3 16-28% 4 >28% 5 0-50 m 1 50-100 m 2 153

Distance to water 100-150 m 3 bodies (streams, 150-200 m 4 rivers, lakes etc) >200 m 5 0-50 m 5 Distance to road (Only 50-100 m 4 highways and main 100-150 m 3 roads) 150-200 m 2 >200 m 1

These scores can be adjusted based on the specific species requirements by assigning weight (Eq. 6.36). For example, some factors such as slope are not applicable for class Aves (birds). Therefore, a weight of zero percent can be applied to remove this factor. And so, in accordance to the site issue or specific species requirement, appropriate weightings for each input layer can be assigned to the final composite layer if applicable. (In this study, equal weight values have been considered because we do not focus on any specific species.)

Habitat suitability (a proxy indicator for habitat connectivity) derives from (Beier et al. 2007):

HS() S W Eq. 6. 36  ii Where HS is habitat suitability, S is the score for factor i, and W is the weight or percentage of importance for factor i.

Then, to obtain habitat connectivity (HC), converts HS value from Eq. 6.36 to an identical range between 0 and 100 percent.

(HS  5) Eq. 6. 37 HC 100 (1 5)

Normalisation: The baseline value is derived from Favilli et al. (2015) and is assigned based on habitat suitability as a proxy indicator as shown in Table 6.26.

Table 6. 26 Normalisation of habitat connectivity (habitat suitability) value

Habitat connectivity (%) Baseline value >75 5 50-75 4 25-50 3 0-25 2 0 1

Improvement strategy/Future scenario testing: According to Figure 6.6, an ideal situation for wildlife conservation would preserve important nodes (core reserves), provide corridors (linkages) between nodes, and establish multiple-use areas (buffer zones) around the nodes and corridors. This pattern satisfies wildlife needs and buffers potential adverse impacts originating from the network. It also provides opportunities for low-intensity human use of the buffer zones around the reserves. 154

Figure 6. 6 Core Reserves, Buffer Zones, and Linkages (Adams and Dove 1989).

Some of the design principles for improving landscape connectivity are: Large patches (nodes) are better than small patches. The chance that wildlife-inhabiting patches will interact becomes disproportionately greater as the distance between patches decreases. In addition, shorter distances between patches, and less contrast between patch and corridor means a higher potential for movement between patches. Continuous corridors are better than fragmented corridors. Wider corridors also work better than narrow corridors. Natural connectivity should be maintained or restored as part of the habitat enhancement strategy. Two or more corridor connections between patches (redundancy) are better than one.

Conservation corridors (links) are an important part of the discussion of biodiversity conservation. If regional- or watershed-scale corridors are impossible or unlikely to succeed, a single, large reserve may be the best choice. In this way, edge and area effects are diminished, population sizes can be larger, and species diversity can be higher, resulting in greater diversity within the ecosystem. If several smaller reserves can be built and connected by corridors, a greater diversity of habitats can be preserved and a larger geographic distribution of populations will be maintained. Separate populations can exist in each reserve, isolated from local disasters affecting survival in the other reserves, while acting as a functional meta- population capable of sustaining the species across the landscape area. The fragmented nature of most agriculturally dominated landscapes suggests that the concept of several small reserves will be most applicable. At the conservation and organisational level, planning and design patches are preferred over reserves. Large reserves tend to support a greater diversity of species. However, if several small patches can be preserved (or created) and connected, the wildlife resource may be equally well served.

155

INDICATOR C6-2: Species diversity

Description: This indicator has been chosen to estimate the potential role of species diversity for the enhancement of the GI network and the conservation of biodiversity and ecosystem services in urban areas. Species diversity is the number of different species, richness and abundance that are represented in a given community. To calculate this indicator, both public green spaces and private gardens that contribute to biodiversity enhancement should be considered. However, in this study, relevant data for species in private and domestic gardens are not available. Therefore, only species diversity in the public domain is calculated.

A variety of methods and equations were developed by scholars to assess species diversity. The most common ones are Simpson and Shannon’s diversity index that has been widely adopted in the ecological studies. In this index, richness and evenness are two main factors that are taken into account when measuring biological diversity: in any given community where richness and evenness both increase, then diversity also increases.

Calculation: To calculate the species diversity (SD), the Shannon’s equitability index was utilized. It can be calculated by dividing the Shannon diversity index by maximum diversity has been widely adopted in the ecological studies as a method to measure species diversity.

Species diversity or Shannon’s equitability score (SD): SDI SD  Eq. 6. 38 ln(S ) Where SD is species diversity or or Shannon’s equitability score (SD), S is is total number of species in the community (richness).

SDI  p ln p Eq. 6. 39  ii

Where SDI is Shannon diversity index, pi is the proportional abundance of the ith species (i) in each grid, ni from the formulas below is the number of individual species type in each grid, N is the total number of species. n Eq. 6. 40 p  i i N

Normalisation: The calculated species diversity accounts for both the abundance and evenness of attributes within the different species types. It typically varies between 0.3-2.5 with a higher index value indicating greater diversity. To ensure consistency with other indicators, it is normalised to a scale ranging between 0.00 and 1.00 (Shannon equitability score) followed by multiplying by 100, where the higher index values indicate higher value diversity (ranked as 100%) within the grid cell. Thus, normalisation of SD value is reported in percent and baseline value, as shown in Table 6.27.

156

Table 6. 27 Normalisation of species diversity value or Shannon’s equitability score

Species diversity (%) Baseline value >80 5 60-80 4 40-60 3 20-40 2 <20 1 Improvement strategy/Future scenario testing: It is obvious that to improve the performance of this indicator, the total area of the public and private gardens in each grid should be increased (noting however that data relating to private gardens were unavailable for this research). This will not happen without public contribution and understanding of their influence in supporting native biodiversity on their own land. Therefore, engaging citizens and land owners is a major challenge for local governments in the support of urban biodiversity. This requires educating residents and modifying attitudes that may prevent the alignment of community behaviours with biodiversity conservation (Parker et al. 2008). Some actions such as workshops, advertisements and educational materials will help to encourage public contributions.

6.3 Health indicators

The healing power of contact with nature is well-acknowledged amongst a variety of disciplines including biology, psychology, sociology and urbanism (Barton and Grant 2006; Coutts et al. 2014; Díaz et al. 2011; Kaplan 1973; Lee and Maheswaran 2011; Macintyre et al. 2002; Ulrich 2002). Accordingly, extensive efforts across disciplines have been made to improve quality of life in neighbourhoods and cities (Barton et al. 2006) through increasing access to natural settings. Barton et al. (2006) proposed an ecological health model which illustrated that the natural environment is fundamental to ecosystem services and human health and well-being.

The World Health Organization defines human health as “a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity” (World Health Organization (WHO) 1946). The Millennium Ecosystem Assessment (MEA) has a broad definition of human health and well-being which includes “material security, personal freedoms, good social relations and physical health” (Millennium Ecosystem Assessment 2003). Diener Ed et al. (1999) stated that human health and well-being not only refers to psychological and physical health but also addresses human feelings and thoughts, life satisfaction, the ability to be self-confident, and processing a sense of happiness and belonging. Hartig et al. (2014) proposed four ways that access to nature may provide health benefits to individuals and communities including: 1) encouraging a more active lifestyle which improves physical well-being; 2) restorative effect (reduction of stress, promotion of mental well-being); 3) social cohesion and increased community interactions; 4) air quality enhancement by plants (which provide additional health benefits). Despite the growing understanding of the health benefits of contact with nature, there is little evidence indicating differential health benefits associated with specific characteristics of GI. 157

This section outlines the key measures or proxy indicators of urban GI that have been used to assess the association of greenery exposure with health and well‐being. The health benefits of GI are categorised here through three indicators (Table 6.28) - physical, social and mental health. In the next section, specific GI characteristics that promote and assist the acquisition of human health within these three categories in addition to their evaluation metrics are proposed. Three proxy indicators have been identified: accessibility by walkable distance; diversity in spatial distribution; and views of greenery.

- Physical health: Physical accessibility (A proximity‐based indicator to any GI types which considers the number of parcels located in a walkable distance (max 400m) of any green space (nodes) within a minimum area of 0.4 hectare). - Social health: Green equity (diversity in spatial distribution of GI). - Mental health: Visual accessibility (View-based exposure to GI).

It is worth mentioning that participants in the online questionnaire scored the physical health indicator as the most important among the sixteen indicators, with mental health achieved as third most important. This emphasises the relevance of measuring health benefits via GI.

Table 6. 28 Health category

GREEN INFRASTRUCTURE ASSESSMENT MATRIX (PERFORMANCE INDEX)

CATEGORY INDICATOR SUB-INDICATOR DESCRIPTION UNIT

C7-1: Calculating travel Improving physical well- Physical cost distance within being (e.g. physical C7 accessibility 300m proximity Score outdoor activity through (proximity in distance to green walking) walkable distance) spaces

Improving social well- C8-1: being (e.g. social GI Equity Shannon diversity HEALTH C8 interaction; social (diversity in the % index INDICATORS integration; community spatial distribution cohesion) of GI services)

Improving mental well- being (e.g. reduced C9-1: Percentage of view depression and anxiety; Visual accessibility C9 to GI in each grid % recovery from stress; (view-based cell attention restoration; exposure) positive emotions)

INDICATOR C7: Improving physical well-being

Description: The characteristics of a built environment can significantly affect a person's decision to be physically active. They can be classified as: density, diversity, proximity, accessibility, connectivity and atmosphere (attractive, comfortable and safe), all of which help 158 to provide a walkable neighbourhood (Austin 2014; Ely and Pitman 2014; World Health Organisation (WHO) 2016). Physical activity in an urban setting can be fostered through planning and physical design. Ding et al. (2011) proposed 12 key factors that influence the use of urban open space as 1) access to and quality of urban parks; 2) recreational facility location and variety; 3) the mix of land uses and desirable destinations; 4) residential density; 5) street connectivity; 6) ease of pedestrian movement; 7) walking and biking facilities; 8) traffic speed and volume; 9) pedestrian safety elements; 10) the degree of neighbourhood civility or order; 11) threat of crime; 12) presence of vegetation.

In addition, aesthetic aspects of natural environments promote physical activity (Ward Thompson and Aspinall 2011). The latest report published by Environment and Planning References Committee (2012) found two particular elements of the built environment that promote healthy lifestyles: provision and quality of green and public open space, and natural environments that encourage active travel (walking and cycling). A large body of studies confirm that accessibility and proximity to green spaces alongside with availability and quality of green space can help to support physical activities (Coombes et al. 2010; Panter et al. 2008).

Connectivity is one of the principles of GI that also supports and facilitates recreational opportunities and physical activities - walking in particular (Coutts 2016). This has been addressed by many studies researching the assumption that more physical activity will reduce obesity (Body Mass Index - BMI) and its associated diseases (Slentz et al. 2005). Ellaway et al. (2005) shows correlation between the amount of greenery in a neighbourhood and the level of physical activity and the consequent self-reported weight loss. Forsyth et al. (2007) examined population density and street connectivity in residential areas and their impact upon children’s BMI. Surprisingly, their study indicates that housing density is not correlated with children’s BMI. However, an increase in amount of greenery and accessibility in a neighbourhood is associated with a reduction of BMI, regardless of residential density. Therefore, environmental factors such as quality of green spaces (Panter et al. 2008), size and type (Brown et al. 2014) , accessibility, exposure and walkable distance (Coutts 2016; Stigsdotter et al. 2010), quantity (Ward Thompson and Aspinall 2011) of GI promote its use for physical activity while simultaneously providing other health benefits. However, age, gender, ethnicity and the perception of safety, are also other influential factors of physical activities and associated health benefits.

Calculation: According to all of the above studies, scholars employed various indicators, such as proximity, size and quality of green spaces as well as access, diversity and residential density, to estimate the influence of natural settings on human physical health. However, all these indicators not only promote physical health but also have parallel impacts on social and mental health. In other words, it is difficult to nominate only one variable that is solely responsible for the positive impacts of GI on physical or social or mental health.

To calculate the performance of GI in increasing physical activity (in this example, walking) accessibility in walkable distance, as a proxy indicator, has been selected by refering to the 159 similar studies (Brownson et al. 2001; Cohen et al. 2007; Evenson et al. 2013; Sallis et al. 2012). Accessibility is defined as the ability of people to reach goods or services as measured by their availability in terms of physical space, affordability and appropriateness (Waters 2016). In addition, accessibility also refers to the provision of services and facilities such as parks, recreational areas and shopping centres (interest points), as well as the means of reaching these facilities in reasonable distance. World Health Organisation (WHO) (2016) proposed a proximity indicator as part of accessibility to encourage more people to walk. This indicator is based on the population density of an area with a 300m walkable linear distance (5 minutes) from an urban park.

Thus, for the purpose of this study and to avoid double counting, a proximity‐based indicator of accessibility, regardless of population density, has been employed to estimate the physical health benefit of GI (walking mode). This proposed indicator is based on proximity of each parcel (residential, commercial etc.) to green spaces without considering the population density. The reason for ignoring population density around destinations (green spaces for this study) is that, if we consider population in the equation, it means that all grid cells with higher population will achieve a higher ranking compared with a low density area (more people have access to the greenery), which is not necessarily accurate in some cases; e.g. increasd population naturally assumes an increase in environmental and social facilities because of an increase in demand. The WHO has recommended that cities provide a minimum of nine square metres of green space per person. However, green space in cities, especially in CBDs, which have high density and high occupation of buildings, is largely related to available area rather than population and demand. This indicates that per capita green space declines in compact cities or highly populated area (Fuller and Gaston 2009).

800 metres’ proximity walking distance to urban facilities was recommended as the maximum baseline value for accessibility suggested by similar studies (Algert et al. 2006; Austin et al. 2005; Witten et al. 2011). However, this walking distance varies based on age group and physical health condition in addition to the destination types. In terms of recreational faclities and open spaces, a 300 - 400 m or five minutes walking distance from a local park (0.4 ha to 1 ha) or 800 m from a neighbourhood open space (1 ha to 5 ha) is suggested (Austin 2014; Rutherford et al. 2013; World Health Organisation (WHO) 2016). It should be noted that while 300 m is a commonly used distance cut‐off (World Health Organisation (WHO) 2016), currently there is no consensus on the walkable proximity distance to green spaces that is linked with physical activity and relaxation‐related health benefits (Ekkel and de Vries 2017).

In this calculation, the linear distance of 300 m was suggested as corresponding to approximately a five minutes walk along walkable pathways regardless of size, type and provision of GI in each neighborhood. Ratio of accessibility in proximity distance is mapped through GIS by Network Analysis and Spatial Analyst extension toolboxes. This procedure simply identifies and ranks grid cells that are located in the service area of each neighbourhood park within a maximum of 5 minutes walkable distance.

160

Normalisation: Normalisation is varied between 1 and 5 percent with five reference value ranges (Table 6.29) indicating the ratio of accessibility to GI in each grid cell.

Table 6. 29 Normalisation of accessibility value Accessibility Baseline value (Proximity in walkable distance) (m) <50 5 50-100 4 100-200 3 200-300 2 >300 1

INDICATOR C8: Improving social well-being

Description: Access and exposure to GI not only benefits individuals but it also may have an impact on public health (Coutts 2016; Mitchell and Popham 2008). Trees and greenery enhance the attractiveness of places for people, in turn promoting social interactions in urban neighbourhoods and a sense of belonging to the place and community, which can lead to reduced crime and increased personal safety (Kim and Kaplan 2004; Kuo et al. 1998; Nieminen et al. 2010). Conversely, insufficient green spaces in a neighbourhood are associated with feelings of loneliness and a lack of social support (Maas et al. 2009; Ward Thompson et al. 2016).

The unequal distribution of green spaces could account for some of the cross-cultural and socioeconomic variations in their use and the consequent amount of social interactions in them. Studies have found that teenagers living in disadvantaged neighbourhoods with poor populations, for example, lack of access to parks and low quality outdoor recreation activities, are less likely to participate in physical and social activities than teens in more affluent neighbourhoods (Babey et al. 2007; Stigsdotter et al. 2010; Van den Berg et al. 2010). Another study conducted by Panter et al. (2008) noted that people in low-income households are more likely to adopt low levels of activity and are least well-served by affordable facilities. Affluent residents, on the other hand, are more likely to live in close proximity to recreational facilities of any type. Differences in physical and social activity are consistent with socioeconomic gradients in many health outcomes and may represent a key pathway through which socioeconomic status affects health. Whilst access to green space appears to be implicitly linked with levels of wealth, what cannot be determined are confounding factors such as individual lifestyles, which could have socioeconomic links. Therefore, green equity, fairness or equal distribution of greenery and other recreational facilities, regardless of differences in poverty, ethno-racial and socioeconomic status of the neighbourhood, is desirable to increase public health (Simon 2016).

161

Calculation: To calculate the diversity in spatial distribution of GI types, the Shannon diversity index was utilised. The Shannon index has been widely adopted in the ecological studies as a method to measure of species diversity. This has been applied to measure and compare landcover diversity index (Willems et al. 2000), social values diversity (Brown and Reed 2012) and the diversity of human activities within the service radius of urban green spaces (Brown et al. 2014) – refer to Table 6.30.

Table 6. 30 Node classifications used in this study adopted from (Brown et al., 2014). Node type (i) Description Typical uses/activities Service Size Radius Mini- Informal recreation spaces, Limited seating and playing Up to 300 m Up to 2 ha park/pocket area, community gardens park Neighborhood Informal recreation spaces Playgrounds and playing fields, 300-800 m 2 – 4 ha park internal trails, seating area, community gardens, dog parks School park Open land associated with Playgrounds and sport field Location Var. schools but open for determined by community use too. school Community Larger park with recreational Playgrounds, sport field and 800m- 2km 12- 20 ha park opportunities that serving recreation centers, picnic and multiple close neighborhoods. seating area, dog parks Large urban Multi-use parks Active recreation areas 2km-5km 100 ha Park Sport park Special use - athletic fields Track and field, soccer etc. Based on Var. location Linear park- Making linkage within and Biking, hiking Trails- location Var. trails between parks variable Nature preserve Heritage landscape, Environmental sustainability Var. (depends 1 ha for each conservation of open spaces emphasis, educational on its size) 1000 population programming increment Regional park Large, easily accessible multi- Multiple activity areas 10 km 500 ha use parks

Shannon diversity index: SDI  p ln p Eq. 6. 41  ii

Where, SDI is Shannon Diversity Index, pi is the proportional abundance of the ith GI type(i) in each grid (considering the coverage of the service radius), ni from formulas below is the number of individual GI type in each grid, N is total number of GI. n Eq. 6. 42 p  i i N Then, equity index of spatial distribution of GI services is calculated by dividing the Shannon diversity index by maximum possible richness in diversity.

SDI Eq. 6. 43 EI 100 ln i Where, EI is equity index, SDI is Shannon Diversity index and lni is natural logarithm of GI services richness that gives us the maximum value for diversity, and i is the number of GI types (frequency is considered as one for each of them - indicating evenness to deliver services). For 162 example, if the study area contains four community parks, two sport fields and five mini-parks then ‘i’ counts as three and ‘N’ counts as 11 in this formula.

Normalisation: The calculated diversity index accounts for both the abundance and evenness of attributes within the different GI types. It typically varies between 1.5-3.5 with a higher index value indicating greater diversity. To ensure consistency with other indicators, it is normalised to a scale ranging between 0.00 and 1.00 followed by multiplying by 100 where higher index values indicate higher value diversity (ranked as 100% equal to 5 in baseline value) within the grid cell. Thus, normalisation is varied between 0 and 100 percent with five reference value ranges (Table 6.31) indicating ratio of distribution of GI in each grid cell.

Table 6. 31 Normalisation of GI equity index value Equity index (%) Baseline value >80 5 60-80 4 40-60 3 20-40 2 <20 1

INDICATOR C9: Improving mental well-being Description: Contact with natural environments can provide restoration from stress and mental fatigue (Coutts 2016; Van den Berg et al. 2010), restore attention and improve cognitive functioning (Kaplan and Kaplan 1989). There are two theories that attempt to explain this: 1) Psychological evolutionary theory proposes that one can reduce stress by having contact with natural settings (Ulrich et al. 1991); and 2) Attention restoration theory suggests that visiting natural settings helps to restore a sense of well-being in those who are suffering from mental fatigue (Kaplan and Kaplan 1989). Both theories are based on the biophilia hypothesis, which argues that humans have an innate need to connect with the natural environment.

A number of studies reported associations between the frequency of visiting green spaces and a variety of psychological, emotional and mental health benefits (Stigsdotter et al. 2010; Ulrich et al. 1991). Ulrich conducted a number of studies in relation to exposure to nature through windows for hospital patients. They revealed that a view of nature through a window has a positive impact on patients, causing them to recover faster from surgery and have a shorter stay at the hospital. Later, he examined stress recovery with passive contact with nature through viewing videotaped images of natural settings. It was found that, as opposed to seeing images of urban settings, “… recovery from stress was much faster and more complete when subjects were exposed to the natural settings…” (Ulrich et al. 1991, 218). However, in contrast to Ulrich’ findings, a number of scholars have shown that having frequent access to, and being in, a natural setting are key principles for stress reduction, over and above just having a view or viewing images (Grahn and Stigsdotter 2003). In addition, increased availability and amount of green space have been shown to improve mental health. Van den Berg et al. (2010) reported that the more green space there is in a 3 km buffer zone around a home, the more stress is

163 relieved. Stigsdotter et al. (2010) shows those who are living further away from a green space (>1km) are more likely to experience higher levels of stress compared with those living near a green space. According to Coutts (2016), effective GI planning should find a balance between built and natural environments so that they deliver mental health benefits.

Generally, it is difficult to measure the non-physical health benefits of GI. More recently, studies have focused on the quality of green spaces, suggesting that the quality might be more important than the amount and quantity of green spaces in the neighbourhood, in terms of impact on mental health (Francis et al. 2012). For example, low-quality green spaces (unmanaged, overgrown) may actually increase anxiety and fear of crime (Kuo et al. 1998). Yu et al. (2016) indicated that view-based exposure to greenery has become an important and much needed task for assessing the quality of urban life. Therefore, to calculate the mental health benefit of GI, a view-based exposure to green area through windows is selected as a proxy indicator.

Calculation: In most studies view to greenery through windows is considered to evaluate the influence of urban vegetation on house price. In addition, the physical characteristics of the built environment (e.g. building heights) intensely impact on GI’s benefits. Yu et al. (2016) proposed the Floor Green View Index (FGVI), which is a three-dimensional assessment method for green visibility from a building. This model puts an observer on all floors and sides of individual building and calculates the average area of the visible green spaces through the windows (Figure 6.8). However, on a larger scale this method is time consuming.

Figure 6. 7 Explanation of Floor Green View Index (Yu et al. 2016).

164

This study proposes a different method to identify the ratio of green visibility. In contrast to the method proposed by Yu et al. (2016), instead of putting observers on all building floors, it is assumed that only one observer is located in the centroid of each GI type and who looks at the surrounded building’s façade. And assumption is if observer can see a proportion of the façade, then anyone in other floors can also see the observer and his/her surrounding area (GI).

ArcGIS was employed to extract all building façades that have view to the GI and the area of the visible building facades in four directions was calculated. Then, this value (m2) is divided by total building facades in the related direction. This value reflects the ratio of average visibility through the buildings’ windows to greenery. This calculation can be more accurate if only the total area of openings (windows) in each façade is included in Eq.6.44 rather than area of visible façade. But this requires more time to collect data.

Vfacade Eq. 6. 44 VA  100

Tfacade

Where, VA is Visual accessibility ratio, Vfacade is area of visible buildings’ façade, and Tfacade is total area of façade in that direction.

Normalisation: Normalisation is varied between 0 and 100 percent with five reference value ranges (Table 6.32) indicating the ratio of visual value of GI in each grid cell.

Table 6. 32 Normalisation of visual accessibility value Visual accessibility (%) Baseline value >80 5 60-80 4 41-60 3 21-40 2 <20 1

165

6.4 Sociocultural indicators

Based on the questionnaire’s results, the sociocultural category consists of three out of the original eight indicators from the initial list: (1) food production; (2) recreation and ecotourism; and (3) improving pedestrian ways (walkability and connectivity).

Table 6. 33 Sociocultural category

GREEN INFRASTRUCTURE ASSESSMENT MATRIX (PERFORMANCE INDEX)

CATEGORY INDICATOR SUB-INDICATOR DESCRIPTION UNIT

Capacity of land to Food production (e.g. produce food urban agriculture; Land productivity through edible % C10 kitchen gardens; capacity gardens (backyards edible landscape and and local community community gardens) gardens) Mixing or distribution of different land uses within a Mixed-use Opportunities for neighbourhood neighbourhood recreation, tourism enhances the (land use C11 and social interaction livability and Score SOCIOCULTURAL composition and (community livability) sustainability of the INDICATORS mixing with GI) neighbourhood and its surroundings as will increase the performance of greenery Improving pedestrian Pedestrian and ways, walkability and cycleway design connectivity C12-1: Walkability criteria (width and Score configurations C12 (e.g. increasing safety; quality of paths; standards) connectivity and linkage with other C12-2: Connectivity Intersection density Score travel modes)

INDICATOR C10: Food production - Domestic gardens

Description: Protection and enhancement of GI is essential in supporting basic ecosystem services such as water, air and the food upon which human health relies. GI plays a vital role in producing food for three ecosystem processes. They are primary production, nutrient cycling and pollination. Primary production is “the synthesis and storage of organic molecules during the growth and reproduction of photosynthetic organisms”(Coutts 2016, 102). 40 % of plants' net primary production is consumed by humans. GI is a very important resource in maintaining the primary production that is currently being lost through land degradation, deforestation and selective overharvesting, which can lead to food shortages. In addition, the production and eventual death and decomposition of plants are fundamental to supporting nutrient cycling as an ecosystem service.

166

GI supports the cycling of nutrients through ecosystems and also provides habitat for organisms that pollinate crops. Pollination is a supporting ecosystem service needed for food production and is dependent on GI. The majority of the world’s crops consumed by humans are moderately to completely dependent on animal-mediated pollination, and reduced landscape and biodiversity endangers the ability of pollinators to do their job (Klein et al. 2007).

Community gardens and home gardens are types of GI, and provide local food production along with other health benefits. People who participate in community gardening not only report receiving higher quality and organic food production, but also improved physical and mental health and increased social interaction (Armstrong 2000). It is not just food itself that can be viewed as medicine but also the activities associated with growing healthy food.

Calculation: Ghosh (2014), in research on nine case studies from Australia and New Zealand, developed a local food energy model for measuring the potential to grow local food, mainly vegetables, in backyards. This model is based on the garden’s size, soil type, total resident population density and total food demand factors. According to Ghosh (2014), the average productive capacity of vegetables in energy units is 1847 kcal per m² of garden per year. This average value, 1847 kcal, is adopted for this calculation.

1847 A Eq. 6. 45 PC  garden 100

Edem

Where, PC is land productivity capacity, Agarden is area of domestic gardens and community gardens in each grid cell, and Edem is the kcal total energy value of annual vegetable demand for the whole population living in that grid cell.

If exact information about the area of the existing home gardens does not exist, substantially, the proportion of required private open space/landscaping area is calculated. In Australia, the Code SEPP (the State Environmental Planning Policy) provides standards with regards to percentage of landscape requirements in each lot based on its size for complying development (Table 6.34). However, this gives us a total required area including front and rear landscaping areas. Usually, 50 percent of the private open space is located in the front yard zone (more hard scape and equipped with seating areas etc) and the rest is dedicated to rear garden and backyards which have deep soil zones and more potential to produce food. Therefore, in Eq. 6.44, only 50 percent of the private open space obtained from Table 6.34 has been considered (rear garden) as potential land to produce food.

Table 6. 34 Code SEPP landscaping requirements Lot size (m2) Landscape requirements (%) <300 m2 10% 300 m2-450 m2 15% 450 m2-600 m2 20% 600 m2- 900 m2 30% 900 m2 – 1500 m2 40% >1500 m2 45%

167

The average energy required daily from vegetables ranges from approximately 12% to 20% of the total daily food intake recommendations. In Australia, recommended daily average food energy intake per person is about 2150 kcal (Haug et al. 2007). The Australian government’s ‘Healthy Eating’ guide also recommends that an average person should intake a daily quantity of vegetables ranging between 425 kcal to 255 kcal. Therefore, the total vegetable demand will be at an average percentage requirement of 15% or about 330 kcal of the recommended total daily diet for an average person (Ghosh 2014).

Yearly energy demand (Edem in kcal) for the total residential population is calculated as:

Eq. 6. 46 Edem  Pop 330  365

Where, Pop is total population per grid cell, 330 is kcal per day (per person), 365 is the conversion factor.

Normalisation: Normalisation is varied between 0 and 100 percent with five reference value ranges (Table 6.35) indicating ratio of productivity capacity to GI in each grid cell.

Table 6. 35 Normalisation of land productivity capacity value Productivity capacity (%) Baseline value >80 5 60-80 4 40-60 3 20-40 2 <20 1 Note that some grid cells may achieve more than 100 percent value. It indicates that area can support a greater number of people than those who live in that cell.

INDICATOR C11: Opportunity for recreation and tourism

Description: Some GI types are specifically designed to provide sports and recreational facilities, including ovals, golf courses and other institutional playing fields. Often, people decide to spend their leisure time based on the characteristics of the environment and recreational possibilities that are geographically closest to them. GI performance would be improved by including an amount of greenery within different land-uses. For example, GI can increase opportunities for recreational activities and ecotourism by connecting the points of interest (amenities) through green corridors. This incorporating and connecting of various land uses can reduce reliance on private cars, encourage walking and cycling (Brown et al. 2009), and increase tourism. The subsequent flow of ecotourism and recreational activities produces additional income and job opportunities for existing residents, which enhances the value of natural amenities in return (Benedict and McMahon 2006).

168

Land use mix (LUM) value as a proxy indicator has been selected to measure the additional value of GI, recreation and ecotourism, through mixing GI with other land uses. LUM refers to a mix of uses within a neighbourhood that enhances the liveability and sustainability of the neighbourhood. To measure LUM, Frank et al. (2006) developed a concept known as entropy score.

Calculation: Entropy score is the most extensively accepted and commonly used method for demonstrating the land-use mix. Entropy generally measures homogeneity of land use in a given study area (Frank et al. 2006). Entropy is expressed as follows:

Eq. 6. 47 ln p j ES  p   j ln N

th Where ES is Entropy score, Pj is the proportion of total land area of j land-use category, N is total number of land uses considered in the study area.

As green surfaces (parks, public and private recreation etc) are part of land-use land cover classification, the assumption is mixing green spaces/recreation land use with other land uses e.g. commercial core and business development as well as medium-high density residential areas. This assumption will increase the performance of greenery by increasing the number of users (visitors) as compared with the performance of isolated GI elements.

Normalisation: Since entropy is normalised using the natural logarithm of the number of land uses (j), its value lies between 0 and 1 where 0 represents homogeneous land use, and 1 indicates that the tract of land is equally distributed across all land use types. Thus, normalisation is varied between 0 and 1 with five reference value ranges (Table 6.36) indicating the ratio of land-use mix in each grid cell. As 100 m by 100 m is a small scale to have all land uses, a cut-off point of 0.5 value for entropy score (ES) has been considered. It means any grid cell that has a value higher than 0.5 ES will achieve a baseline value of 5 (the highest rank).

Table 6. 36 Normalisation of land use mixed value

Entropy Score- LUM (score) Baseline value >0.50 5 0.40-0.50 4 0.30-0.40 3 0.20-0.30 2 <0.20 1

169

INDICATOR C12: Improving pedestrian ways, walkability and connectivity (Design criteria): Improving the quality of pathways encourages people to walk and cycle. A good street network design works as a platform and structure for GI planning. To measure this indicator, two proxy sub-indicators have been considered which are Walkability and Connectivity scores.

INDICATOR C12-1: Walkability (design criteria)

Description: Including elements of greenery along footpaths and cycleways can provide safer and more convenient pathways for pedestrians and cyclists. The width of the path is an important factor in providing greater opportunity for street tree planting and linear green verge, which in turn dictates the level of walkability of an area, helping to solve potential conflicts between usage and safety. The Walkability Score is an indicator that indicates the degree to which a corridor is regulated with standard linear greenery alongside pedestrian and cycleways that made it attractive and safe to walk and cycle. Site walkability is assessed with this indicator by looking at configurations and arrangements of green infrastructure elements that retrofit into the cycleway and footpath. For example, GI can provide shade to paths or safety by separating cycleways, footpaths and driveways with linear green verges.

Calculation: To calculate Walkability (as one of the indicators that improves pathways in accordance with GI planning), scores were assigned based on design criteria. These scores indicate whether a given corridor achieves these design criteria or not. Figure 6.8 shows the acceptable standard for walkable and green street design.

Figure 6. 8 Typical cross-section of a separated protected bicycle lane(schematic diagram is adopted from Levasseur (2014, 4))

170

Tables 6.37, 6.38 and 6.39 show desirable widths and acceptable ranges of width for shared path, and two-way and one-way separated paths, respectively, provided by Austroads (Levasseur 2014).

Table 6. 37 Shared path (Levasseur 2014, 94)

Shared path standard width Path width (m) Local access path Commuter path Recreational path Desirable minimum width 2.5 3.0 3.5 Minimum width – typical maximum 2.5-3.0 2.5-4.0 3.0-4.0

Table 6. 38 Separate two-way path widths (Levasseur 2014, 95)

Separate two-way path standard Path width (m) width Bicycle path Footpath Total Desirable minimum width 2.5 2.0 4.5 Minimum width – typical maximum 2.0-3.0 ≥1.5 ≥4.5

Table 6. 39 Separate one-way path widths (Levasseur 2014, 95).

Separate one-way path standard Path width (m) width Bicycle path Footpath Total Desirable minimum width 1.5 1.5 3.0 Minimum width – typical maximum 1.2-2.0 ≥1.2 ≥3.4

Normalisation: Normalisation is achieved by classifying and assigning weight (1 to 5) to the cycleways and footpaths based on the existence of green infrastructure elements in their design criteria (Table 6.40) (Dizdaroglu 2013).

Footpath = 1 point Vegetative buffer zone = 2 points Mixed traffic cycleway or shared path = 1 point Separated cycleway = 1 point Separation strip = 1 point

Table 6. 40 Normalisation of walkability score (Dizdaroglu 2013)

Walkability design score Baseline value Footpath + Vegetative buffer zone+ Separated cycleway+ Separation strip 5 Footpath + Vegetative buffer zone + Mixed traffic cycleway (cycle route pavement markings) 4 Footpath + Vegetative buffer zone 3 Shared footpath 2 None 1

171

INDICATOR C12-2: Connectivity

Description: Parks need to be “linked to one another and to surround residential neighbourhoods” (Austin 2014). A key principle of the GI strategy is related to network connectivity and integration. Connectivity is defined as the degree to which roads, pedestrian walkways and trails are connected. A well-connected network is expected to ease travel between places (Agampatian 2014) and offer shorter and more alternative routes (Frank et al. 2012). This encourages recreational activities and active travel modes (where people decide to walk or cycle rather than use a car) by way of better urban GI planning and design.

The key purpose this indicator aims to achieve is related to street structure as a platform for greenery. Figure 6.9 shows two different streets plans and their level of connectivity. Based on these two street structures, it is obvious that increasing intersection density and decreasing dead-ends is critical to increasing connectivity and encouraging walking. The figure on the left provides longer paths with more opportunity for GI planting alongside the footpath, however, the figure on the right provides a more effective and efficient platform for GI planting. This design platform will increase performance of GI. The shortest path provides more alternative paths, encouraging people to walk rather than forcing individuals to take only one existing footpath, which may not be practical.

Low Connectivity High Connectivity

Longer route, only one route option Direct route, multiple route options

Figure 6. 9 Comparison between low and high connected network between two points (Frank et al., 2003) Calculation: Connectivity makes existing and future GI planning more effective and efficient by increasing the probability of a greater number of users. The GI measurement of connectivity is basically related to the physical characteristics and design criteria of the network as its base. Increasing intersection density as one of the physical characteristics of the network provides shorter and alternative pedestrian pathways (Frank et al. 2012). This indicator also identifies corridors that may be more or less favourable for walking, as well as planting more trees alongside these links in future development scenarios.

Therefore, intersection density, as a proxy sub-indicator for connectivity, is measured by summing the number of intersections (excluding dead-ends) and analysing the results through ArcGIS using the Hot Spot Analysis extension tool to calculate the z_score (normalisation procedure). This tool clusters intersections spatially in relation to the value of neighbouring 172 intersections. The result indicates that intersections with a high and desirable value are statistically significant when surrounded by high number of intersections (higher level of connectivity).

Normalisation: The average z_score value in each grid cell is reclassified using natural break classification based on a scale of 1-5 (Table 6.41), indicating the ratio of connectivity in each grid cell. Grids with high value are preferred over networks that include more intersections and less cul-de-sacs as well as shorter block length, thus decreasing distances between destinations (e.g. parks) and encouraging walking and bicycling.

Table 6. 41 Normalisation of intersection density value

Connectivity (intersection density) Baseline value High 5 Medium-high 4 Medium 3 Medium-low 2 low 1

6.5 Economic indicators

Due to the multifunctional nature of GI, it can be difficult to monetarise its benefits, as different functions may need different forms of units and measurements and some of the values provided by GI remain difficult to quantify in monetary terms, particularly those associated with culture and aesthetics. However, attempting to place a monetary value on the key functions performed by GI is useful in communicating with stakeholders and the community and can feed directly into the policy decision-making process (Vandermeulen et al. 2011). Economic evaluation has been widely discussed in the literature related to GI service monetary values for atmospheric carbon dioxide reduction, energy reduction, stormwater runoff reduction and air pollution mitigation, as well as property value assessment (Akbari 2002; Brack 2002; Coutts 2016; Maco and McPherson 2003; McPherson and Simpson 1999; Naumann 2011; Nowak et al. 2006).

In the subsection of this study that relates to the economic evaluation of GI, interview participants selected four indicators out of ten from the initial set of indicators, including: the value of avoiding CO2 emissions and carbon sequestration; avoiding energy consumption from cooling and heating; air pollution removal; and reducing the cost of using private cars by encouraging people to walk and cycle (Table 6.42). Three of these selected indicators are from the ecological category that have been calculated by iTree Eco. Therefore, the monetary value that was used in the tool is applied here too. The value of the fourth indicator was proposed to monetarise based on value for changing travel mode from passive to active in accordance with the provision of green and high quality pathway and cycle way (per kilometre).

This economic category is not included in the GIS mapping, because the indicators have been mapped through other categories: the only difference here is monetarising the value that is more

173

understandable by stakeholders and communities. This set of four key indicators can give the decision-makers an overall view on the benefits of existing and proposed GI planning. The next sub-section explains these four indicators and unit cost (Australian dollar value) per indicator.

Table 6. 42 Economic category GREEN INFRASTRUCTURE ASSESSMENT MATRIX (PERFORMANCE INDEX)

CATEGORY INDICATOR SUB-INDICATOR DESCRIPTION UNIT

In Australia (iTree Value of avoided CO2 Eco) was calculated Dollar value per C13 emissions and carbon for 2016 at $24.80 $ tonne carbon sequestration (AUD) per tonne carbon (metric).

Value of avoided energy consumption Dollar value per $0.037 (AUD) /Kwh (e.g. reduced demands C14 Kwh and MBTU and $2.970 (AUD) $ for cooling and /MBTU heating)

O3 and NO2 = $673 per metric tonne, CO Dollar value per = $24.15 per metric Value of air pollutant metric tonne of tonne, PM10 and C15 $ removal/avoidance individual pollutants PM2.5= $185 per removed by GI metric tonne and SO2 = $471 per metric tonne. ECONOMIC INDICATORS Length of green sidewalks and/or cycle ways in kilometers multiplying by $0.334 (AUD)×365 This monetary value Reducing cost of using comprises four avoided costs (per private car by Avoid cost by day): $0.207/km increasing walking and shifting to active C16 walked or cycled + $ cycling (e.g. shifting travel mode (per $0.059 travel mode) alongside year) green paths environmental benefits of reducing noise, air quality and GHGs emissions+ $0.052/km avoided cost of infrastructure provision+ $0.016/km avoided parking cost

INDICATOR C13: Value of avoided CO2 emissions and carbon

sequestration

Indicator 3 in the ecological category discussed GI's, and in particular trees', ability to capture and store carbon from the greenhouse gas CO2 during the photosynthesis process. Pakzad et al. (2015) reviewed existing tools to estimate carbon storage and sequestration performed by GI

174 and the i-Tree application was recommended as the best tool. It was developed in the US and it places dollar values on a number of GI benefits. The iTree model estimates the value of the carbon stored and sequestered annually through biomass equations, multiplied by unit cost. The model uses the estimated marginal social cost of carbon dioxide based on the carbon value for the United States Interagency Working Group (U.S. Environmental Protection Agency 2015) and then updated to current dollar values. Currently, the value used in Australia (with i- Tree Eco) was calculated for 2016 at $24.80 (AUD) per tonne carbon (metric).

INDICATOR C14: Value of avoided energy consumption

Studies confirm that green roofs, street trees and increased urban green spaces have the effect of making individual buildings more energy efficient by reducing heating and cooling demands. For example, street trees, according to Heisler (1986), reduce cooling costs by 20-50% and heating costs by 10-15% for residential allotments. McPherson and Simpson (1999) developed a guideline to calculate energy benefits of trees based on a series of lookup tables. This calculation requires input variables for the location and structure of trees and then dollar values are calculated based on energy costs of the location. The iTree tool also calculates the monetary value of trees' energy saving based on this guideline and, similarly, the U.S. CUFR Tree Carbon Calculator (CTCC). Electricity and MBTU values in iTree Eco in Australia are based on $0.037/Kwh and $2.970/MBTU. Tree canopies less than 6m height or further than 18m away from buildings are removed from the calculation model.

INDICATOR C15: Value of air pollutant removal/avoidance

The iTree application also can estimate hourly pollution removal and its dollar value based on local weather and solar radiation data, pollution data, leaf area index, leaf-on and leaf-off dates, and geographical factors (Nowak et al. 2006). The value attributed to pollution removal by trees is estimated within the model using the median externality values for the USA for each pollutant. These values are currently adjusted for Australia by Melbourne Urban Forest and Arboriculture Australia to $ per metric tonne as O3 and NO2 = $673 per metric tonne, CO = $24.15 per metric tonne, PM2.5= $185 per metric tonne and SO2 = $471 per metric tonne. These values are considered to be the estimated cost of pollution to society that is not accounted for in the marketplace for the goods or services that produced the pollution, e.g. the costs that are involved for a source polluter (e.g. power station) to install, maintain and relace technology that removes these pollutants from their business activities.

INDICATOR C16: Reducing cost of using private cars by increasing walking and cycling alongside green paths

Studies show that improving urban quality of life to promote pedestrian activity will have a small but significant positive impact on businesses. GI can also play an important role in improving the quality of cities by creating attractive places to live with welcoming images, and providing settings for a range of events and activities that can boost the local economy. A study 175 by Whitehead et al. (2006) in Manchester, England demonstrates the positive impact of general urban quality improvements, such as walkable places, on economic activity. This indicates a significant boost in willingness to shop, do business or work in areas before and after ‘walkability’ improvements; this was associated with a 20-40% increase in foot traffic and a 22% increase in rents in pedestrianised retail areas.

High quality footpaths and cycleways can increase active travel modes. A mode shift towards active travel can reduce traffic congestion and its associated costs. The Australian Government (2013) provides a dollar value per day per kilometre of footpath and/or cycleway for reducing the cost of using private cars by increasing walking and cycling. This monetary value comprises four avoided costs (per day): $0.207/km walked or cycled (decongestion) + $0.059 environmental benefits of reducing noise, air quality and GHGs emissions+ $0.052/km avoided cost of new infrastructure provision+ $0.016/km parking cost savings (Auditor-General of Victoria 2013; Australian Government 2013). Therefore, to monetarise this indicator, the length of green footpaths and/or cycleways in kilometres is multiplied by $121.91 yearly benefit ($0.334 (per day) × 365).

Summary

Analysis from the questionnaire of experts identified 16 key indicators within four categories (ecological, health, sociocultural and economic), reduced from the original set of 30 indicators, as being important for measuring the sustainability performance of GI. This set of 16 indicators established the structure for the assessment model. In this chapter, descriptions, equations and baseline values for each of the sixteen key indicators and sub-indicators have been described. In the next chapter, the proposed indicator-based model has been calibrated through a pilot study of Parramatta CBD - a major urban area in the geographic heart of the Sydney metropolitan region.

176

CHAPTER 7: Case study application of the model

Introduction to the case study

In this chapter, the proposed indicator-based model has been calibrated through a pilot study of Parramatta CBD. This study area has good potential in testing the model because of its urban form and structure as well as data availability. In addition, Parramatta City Council is currently developing detailed greenspace planning for this part of the city, called the ‘Green Grid Master Plan’, that will incorporate green infrastructure as a part of its design strategy. In the testing, analysing and building of the proposed model developed for this research, a number of software programs have been employed - ENVI, ArcGIS and iTree Eco. The visualisations of the results are presented at the end of this chapter.

This chapter is organised into six sections. Following this introduction, the next section provides a description of the case study that will serve as a foundation for developing the measurement matrix. Section 2 describes data acquisition and preparation following by the methodology of data analysis in section 3. Section 4 presents the results for the 16 indicators in the study area. Section 5 outlines the aggregation of the results. It shows the overall level of sustainability of the study area. Finally, the chapter concludes with a summary and discussion.

7.1 General characteristics of the study area

Parramatta CBD lies in the central-western area of Sydney, Australia, and has been selected as a basis to generate a normalisation and benchmarking procedure for the indicator set. It is located on the banks of the Parramatta River (33°48′54″S and 151°00′4″E). Parramatta is a major business and commercial centre and has the second largest CBD in the State of New South Wales (Rauscher and Momtaz 2016) - the sixth largest in Australia. The local government area of Parramatta spans 61.33 square kilometers, stretching to the local government areas of the Hills and Hornsby in the North, to Auburn in the South and South East, and from Holroyd in the West to Ryde in the East. Parramatta CBD is about 2.6 square kilometers (Figure 7.1).

The selection of this study area is based on three criteria. First, the Sydney metropolitan strategy 2014 proposed to establish the Green Grid vision for Parramatta area in order to shape and support sustainable development while increasing quality of life and livability (The NSW Government Planning and Environment 2014). In this regard, the Government Architect’s Office (GAO) proposed the establishment of a network of green areas for Parramatta. This project determined where additional open space was required in order to ensure equal access to walking and cycling (Government Architect’s Office NSW Public Works 2016). Second, this study area represents mixed land use and a high density of commercial and residential buildings with a variety of green infrastructure types and potential for additional green spaces. Thirdly, digital spatial data and tree inventory are available for this research project. 177

Figure 7. 1 Parramatta local government area (LGA), CBD and green infrastructure distribution.

Parramatta has a temperate, subtropical climate with warm to hot summers and moderate to cool winters with an annual average minimum temperature of 12.2 °C and maximum of 23.3 °C. While the average air temperature in January reaches up to 28.4°C, sometimes temperatures reach more than 40 °C during heatwaves; the average winter minimum reaches 6.2 °C. Average annual rainfall is 970.6 mm, with highest rainfall occurring between January and March. Table 7.1 summarises the weather statistics for Parramatta.

Table 7. 1 Summary of monthly weather statistics in Parramatta (Australian Bureau of Statistics 2017)

Statistic Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Ann Years ual Mean max. 28.4 27.8 26.3 23.8 20.6 17.8 17.4 19.1 21.7 24.0 25.4 27.4 23.3 1967- temperature 2016 Mean min. 17.6 17.6 15.8 12.9 9.9 7.6 6.2 7.1 9.3 12.0 14.1 16.2 12.2 1965- temperature 2016 Mean rainfall (mm) 105. 121. 106. 93.6 69.5 91.2 46.2 56.9 52.4 67.7 86.6 73.6 970. 1967- 7 2 7 6 2010 Mean 9am wind speed 7.3 6.4 6.4 6.7 6.7 7.2 7.7 9.1 9.8 9.8 8.4 8.1 7.8 1967- (km/h) 2010 Mean 3pm wind speed 14.5 13.0 12.2 10.8 9.3 10.4 10.6 13.2 15.2 14.9 15.6 15.4 12.9 1967- (km/h) 2010 Mean 9am relative 74 79 80 76 79 78 75 67 63 62 69 70 73 1987- humidity (%) 2010 Mean 3pm relative 57 59 59 58 60 59 55 46 46 49 54 55 55 1987- humidity (%) 2010

178

Average wind speeds range from 9.3-15.6 km/h (at the height of 10 m above the ground surface). In 2016, the maximum daily summer solar radiation flux density peaked at around 32.3 kWh/m2, while, in winter, the maximum peak was somewhat lower at around 11.1 kWh/m2 (at Parramatta North, Masons Drive station). Table 7.2 and Figure 7.2 show the monthly prevailing wind direction in Sydney and annual wind rose for Parramatta, respectively. The wind statistics are based on real observations from the Bureau of Meteorology weather station at Parramatta.

Table 7. 2 Monthly prevailing wind wave direction in Sydney (Direct NE, Prevent S and W)

Months Prevailing Wind Direction North-Easterly Southerly Westerly Jan-March x x April x x May-Aug x Sep x x Oct-Dec x x

Figure 7. 2 Annual wind rose for Parramatta

For the purpose of this research, an area of 2.83 square kilometers with about 500 metres of buffer area around the Parramatta CBD has been selected. It consists of 54.06% pervious, 4.93% water body and 41.01% impervious surfaces. Figure 7.3 demonstrates eight land surface classifications for the study area. As part of the methodology, this classification has been done through ENVI, version 5.1, image-processing software.

179

Figure 7. 3 Land surface classification of study area (pervious and impervious).

In general, the urban structure in Parramatta CBD is compact and has a diversity of land uses, with narrow to wide streets and a variety of canyon orientations. Based on the results from unpublished local climate zone classifications by Dr Graciela Metternicht and Yueyi Feng at Univeristy of New South Wales, this study area has been divided into 15 differentiated climatic zones based on the Stewart and Oke (2012) classification. The Local Climate Zone (LCZ) classification scheme was initially developed by Stewart and Oke (2012) comprises 17 zones based mainly on characteristics of surface structure (e.g., building and tree height & density) and surface cover (pervious vs. impervious). This standardised classification is a logical starting point that helps to unifying research methods, findings and documentations especially in climatological studies among scholars. An overall view of this 15 climate zone classifications for the study area is summarised in Figure 7.4.

180

Figure 7. 4 Local climate zone classifications based on Stewart and Oke (2012)

Sites were characterised by differences in the aspect ratio (H/W) of street canyons, street orientation, surface material properties and GI types. A more detailed description of the site’s mean aspect ratio analysis of canyons is given in Figure 7.5. In this analysis of height-to width, dominant height alongside each canyon between building and tree canopy has been considered based on LCZs classifications. Figure 7.5 shows typical cross-sectional canyons in the study area.

Figure 7. 5 Typical cross-sectional canyons

Figure 7.6 illustrates the urban geometry analysis of the study area. It indicates that more deep and narrow canyons are located in the centre and south-east of the study area, while open and wide canyons with open medium and low-rise buildings are located in the north of the site. 181

Figure 7. 6 Mean height-to-width ratio of street canyons classifications.

The study area comprises 1704 land parcels with an average parcel size of 1115.84 m2. Of the 1704 parcels, 7.7%, 2.3% and 38.6% are commercial, business development and mixed land use areas, respectively, 7.3%, 2.7% and 14.6% low, medium and high density residential zones, 6.9% infrastructure, 1.7% natural waterway, 1.3% neighbourhood centre, 3.3% and 11.2% private and public recreational zones respectively, while 2% is dedicated to recreational waterway and 0.4% special activities such as places of worship (Table 7. 3 and Figure 7.7).

Table 7. 3 Percentage of individual land uses in study area (Results derived from GIS data analysis)

Land uses Percentage of land use Commercial 7.7% Business development 2.3% Mixed land use 38.6% Low density residential zone 7.3% Medium density residential zone 2.7% High density residential zone 14.6% Infrastructure 6.9% Natural waterway 1.7% Neighborhood centre 1.3% Private recreational zone 3.3% Public recreational zone 11.2% Recreational waterway 2% Special activities e.g. place of worship 0.4% 182

Figure 7. 7 Land use zoning classifications

Mesh blocks data is the smallest area for population distribution that the Australian Bureau of Statistics (ABS) collects and publishes; the latest available data is from 2011. The case study area consists of 286 mesh blocks that each contain approximately 30 to 60 dwellings. About 17,364 people live in the study area, where there is an average population density of 61.3 person per hectare based on mesh blocks data (Australian Bureau of Statistics 2011). Figure 7.8 illustrates the population density distribution (residents) which was normalised based on area of each mesh block.

183

Figure 7. 8 Distribution of population density (Residents)

In terms of GI type, Parramatta Park, Ollie Webb Reserve, Elizabeth Farm, Doyle Ground and Belmore Parch are the major green spaces that are located within and around the study area. Figure 7.9 shows the distribution of existing patches (nodes) and green corridors (links) in study area.

184

Figure 7. 9 GI distribution (links and nodes)

7.2 Green structure of the study area

The overall tree density in the study area is 1,780 trees that represents 5% of the total urban trees in Parramatta. This provides 0.248 square kilometers of leaf area. Table 7.4 summarises the ten most common tree genera within the study area. The most dominant genera, in terms of leaf area, are Lophostemon (24.3%), Platanus (10.7%), and Callistemon (10.0%). The results have been obtained from an analysis by iTree Eco.

Table 7. 4 Ten most dominant tree genera in study area.

Percent Percent Genera Population Leaf Area Lophostemon 24.3 24.1 Platanus 10.7 20.6 Callistemon 10.0 5.6 Jacaranda 8.4 6.8 Liquidambar 3.4 8.7 Corymbia 4.4 5.0 Eucalyptus 3.4 5.9 Flindersia 7.2 1.2 Melaleuca 3.2 4.3 Pyrus 4.0 0.2

185

Figure 7.10 presents tree population in the study area by diameter at breast height (DBH is the stem diameter of the tree at 1.4 meters vertically above the base of the tree). Trees with a trunk diameter of less than 30 cm constitute more than 50% of the tree population.

Figure 7. 10 Percentage of tree population by diameter at breast height (DBH).

7.3 Data acquisition and preparation In order to develop the model and apply the selected indicators, GIS data with detailed specifications for required information for each indicator, were formally requested from the Parramatta City Council. In addition to the GIS layers, a street tree database, provided by Parramatta City Council in May 2015, was acquired. This database helps with accuracy of the model validation and facilitates the calculation of each indicator that was used as an input to the iTree Eco tool (v.6) and ArcGIS.

Having reviewed Parramatta City Council documents and their applications, it can be concluded that the Council is moving ahead in the direction of applying GI planning concepts with emphasis on walkability and making streets, parks and open spaces desirable and sustainable places to live near. The most recent documents are Parramatta Way (2016) and Public Domain Guidelines (2016). Additional published documents by the Council that were considered in conjunction with applying indicators in this study area are: Planning proposal- Parramatta CBD (2015); Parramatta Biodiversity 2015-2025 (2015); Parramatta City Centre Street Tree Masterplan; Parramatta 2038 - community strategic plan (2013); Design Parramatta (2012); Parramatta local environment plan (2011); The Parramatta City Centre Street Tree Masterplan (2011); Parramatta Street Tree Plan (2011); Development Control Plans (2011); and Street Tree Design Guidelines (2008).

186

7.4 Methodology of data analysis The expert respondents identified 16 key indicators as being significant and worth including in the indicator-based assessment model for measuring the sustainability of green infrastructure, as discussed in Chapter 4. The previous chapter established the measurement scale and identified equations to quantify each of these indicators. Table 7.5 gives a summary of equations, associated variables and required input data assigned to each indicator in the study area.

187

Table 7. 5 Summary of equations, variables and baseline value level

Applicability Indicator Sub-indicator Equation Variables (required data) Unit level of baseline value 2 ()A  Ai = area of each surface cover (m ) i ii ESA 100 αi = average albedo value of each surface C1-1: Effective surface  Ai type % International albedo ESA= effective surface albedo ratio (%)

ET0 = Reference evapotranspiration

AGi = The area of green cover in each grid ()AKKK   cell C1-2:  Gi si di mci ET ET  100 % International Evapotranspiration rate L 0 A Ks = species coefficient  G i Kd = density coefficient ETL = landscape evapotranspiration C1: Climatic Kmc = microclimate coefficient and Canyon aspect ratio (H/W) microclimatic H= average canyon height (tree canopy or modifications  H VEN  canyon& Orientation building height) C1-3: Ventilation  W Score Site specific Green urban geometry W=average canyon width VEN= ventilation score Street orientations (N-S,E-W,NE-SW,NW- SE) NORMSDIST= normalisation procedure from iTree results, X LA   XLA = sum of LA of all individual trees in International (hot SE NORMSDIST *100 C1-4: Shading effect  each grid % and humid climate) SE= shading effect µ = arithmetic mean of total LA σ = standard deviation of total LA NORMSDIST= normalisation procedure X   C2: Air quality C2: Air Pollutant APR NORMSDIST AP *100 from iTree results,  % International improvement removal  XAP = sum of air pollutions removed by APR= air pollutant removal (%) individual trees in each grid (gr per year) 188

µ = arithmetic mean of total air pollutions removed σ = standard deviation of total air pollutions removed

(,)ik C = carbon sequestration value for individual tree in each grid that was calculated by iTree Eco platform Nmax/Grid = maximum number of tree per grid.

(,)ik C Alternative equations bellow can be used if i/ Grid C3: Carbon C3: Carbon storage & CSR 100 21.77 N the iTree wasn’t adopted in the study area % International offsets sequestration max/Grid or if allometric equation not available for CSR= carbon sequestration ratio the specific species. 2.310647 BroadleafBiomass() DW 0.56  0.280258  ( dbh ) 2.580671 ConiferBiomass() DW 0.48  0.05654  ( dbh ) CO2 Biomass ( DW )  1.28  0.50  3.67  0.80

dbh = diameter at breast height (1.4 m) NORMSDIST= normalisation procedure from iTree results, Xcool = total cooling benefits (kwh) C4: Reduced µ = arithmetic mean of total cooling X   benefits building cool ESC  NORMSDIST *100 energy used C4: Energy saving  σ = standard deviation of total cooling % International for cooling and ESC= Energy saving for cooling benefits heating Important variables: Tree azimuth (direction) Distance x ≤18.3m Tree height L≥6.2m

LP = Annual pollutant load (kg) in each grid C5: WQI L (1  R )  R  % A cell Hydrological C5-1: Water quality  p GI(n 1) GI ( n ) GI ( n ) 2 % National AGI = area of each GI type n (m ) regulation n WQI = Water quality improvements ratio 189

RGI = relative percentage of pollutant 6 reduction by each GI type n (%) LPPRCA(     )  10 P J vu u u P = Precipitation, mm/year u PJ = Ratio of storms producing runoff LP = Annual pollutant load (kg) in each grid cell (default = 0.9) RI0.05  (0.009  ) vu u Rvu= Runoff Coefficient for land use type u Rvu = Runoff coefficient for land use type u Cu = Even Mean Concentration (EMC) of pollutant in urban runoff for each land/ land use type cover (mg/litre) 2 Au = Area of land use type u (m ) Iu = percent of impervious cover

iTree Eco: In = infiltration to the pervious surface, P In  S  R  E S = depression storage on both impervious and pervious surface, C5-2: Avoid surface P = precipitation % International runoff R = runoff from impervious cover,

Ravoided  I  In  S E = evaporation,

Aavoided = annual avoided runoff volume I = rainfall interception depth HS() S W Si = score for factor I in resistance surface  ii layer, HS = habitat suitability C6-1: Promoting W = weight or percentage of importance for conservation (habitat % International (HS  5) factor i connectivity/suitability) HC 100 (1 5) C6: HC= habitat connectivity Biodiversity- SDI pi = proportional abundance of the ith SD% 100 species(i) in each grid, protection and ln(S ) enhancement ni = the number of individual species type C6-2: Species diversity SD = Species diversity or Shannon equitability score in each grid, score (equitability % International SDI  p ln p N = total number of species, score)  ii n S= total number of species in the p  i i N community (richness), SDI= Shannon diversity Index 190

Cost = average travel cost distance value between each cell and nearest green space

C7: Improving PA average(Cos tDistan ce ) Creating a travel cost surface within 300 physical well- C7: Physical metre walkable distance (or 5 minutes) of Score International accessibility (Proximity) being PA = Physical accessibility every green space and then calculate distance from each grid cell walking across nearest footpath. SDI pi = the proportional abundance of the ith GI EI 100 type(i) in each grid (considering the ln i coverage of the service radius),

C8: Improving ni = number of individual GI type in each SDI  p ln p grid, social well- C8: Equity  ii % International being n N = total number of GI types (with their p  i i N frequency) EI = equity index i= number of GI types (frequency is equal SDI = Shannon diversity index to one)

Vfacade = area of visible buildings’ façade, C9: Improving Vfacade VA  100 mental well- C9: Visual accessibility Tfacade = total area of façade in that direction % International Tfacade being VA is Visual accessibility ratio

Agarden = area of domestic gardens and 1847 Agarden PC 100 community gardens in each grid cell C10: Food C10: Land productivity Edem Edem = the kcal total energy value of % International production capacity E Pop 330  365 annual vegetable demand for the whole dem population living in that grid cell PC is land productivity capacity th C11: Pj = proportion of total land area of j land-use category, Opportunities ln p for recreation, C11: Mixed-use ES  p  j N = total number of land uses considered  j in the study area Score International ecotourism and neighbourhood ln N social ES is Entropy score interaction

191

Footpath = 1 point C12: Vegetative buffer zone = 2 points Mixed traffic cycleway or shared path = 1 Improving C12-1: Walkability Score National pedestrian point ways, Separated cycleway = 1 points walkability Separation strip = 1 point and connectivity Walkability design score Nint = total number of intersections in each ID N C12-2: Connectivity  int grid cell Score International ID is intersection density C13: Value of avoided CO2 In Australia (iTree Eco) was calculated for 2016 at $24.80 emissions and C13: Dollar value per - $ National tonne carbon (AUD) per tonne carbon (metric). carbon sequestration C14: Value of C14: Dollar value per avoided energy Kwh and MBTU $0.037 (AUD) /Kwh and $2.970 (AUD) /MBTU - $ National consumption C15: Value of C15; Dollar value per O3 and NO2 = $673 per metric tonne, CO = $24.15 per metric air pollutant metric tonne of tonne, PM10 and PM2.5= $185 per metric tonne and SO2 = - $ National removal/avoid individual pollutants $471 per metric tonne. ance removed by GI C16: Reducing Length of green sidewalks and/or cycle ways in kilometers cost of using multiplying by $121.91 (AUD) private car by C16: Avoid cost by This monetary value comprises of three avoided costs (per shifting to active travel - $ National increasing day): $0.207/km walked or cycled +$0.059 environmental mode (per year) benefits of reducing noise and GHGs emissions+ $0.052/km walking and avoided cost of infrastructure provision+ $0.016/km avoided cycling parking cost

192

After data preparation, the study area was divided into 100 × 100 metre grid cells with the ArcGIS software. Then, each indicator was calculated separately in each grid and the results were ranged normalised and ranked between 1 and 5 (low to high), indicating comparative levels of sustainability baseline performance. GIS distinguishes each baseline value level by color ranges, similar to a heatmap (red to green). At the next stage, the spatial aggregation was determined to combine the multidimensional indicator scores into a single composite index. There are various methods of aggregation proposed by Nardo et al. (2005). However, the arithmetic mean formula was decided upon to aggregate scores from all indicators and visualize the final results in each grid in five comparative sustainability levels between 1 and 5. Figure 7.11 illustrates the spatial aggregation of the surface albedo sub-indicator as an example. As shown in this example, the percentage of each land surface type is multiplied by its albedo coefficient and then summed into a single composite score for each grid cell.

Figure 7. 11 Spatial aggrigation of the surface albedo

The final aggregation results demonstrate the level of GI performance in individual grid cells (weak or strong performance). This helps policy makers and designers identify the parts of the study area that require special design attention in order to improve performance and track which type of indicator is not performing well. For example, if one of the grids is coloured yellow, a designer can trace back and identify an ecological category, for example, energy and runoff reduction, which did not achieve a good score. They will then be able to direct their proposed design intervention accordingly. For example, a designer could choose to plant more appropriate tree species for the goal in mind, with the appropriate direction and distance away from buildings. In terms of another indicator that does not perform - runoff reduction - a designer can work on designing an engineered GI type such as a rain garden or changing the percentage of impervious to pervious surfaces (selecting surface cover with lower runoff coefficient) and propose to plant trees with a wider crown canopy that increases the level of interception during precipitation.

193

This indicator-based model can also evaluate various design scenarios and the final results can be compared with the current condition. At the end of each scenario testing process, the four proposed indicators in the economic category can assist in the monetary value comparison of each alternative in order to make it easier to choose the best multidimensional performance value.

This model also benefits local city councils; it can be a useful guide for pre-tender processes for identifying areas that require new landscape/urban design, helping to determine design requirements for specifically identified grid cells. This will provide economic benefits because it will preemptively address an issue that would be a separate part of the project budget.

This proposed model serves as a multiscale assessment tool, which can measure the performance of GI for not only a single tree, shrub or one square metre of green groundcover, but also for a wide network of greenery on the macro scale. Because this model uses iTree Eco software to analyse the ecological performance of urban greenery and can export these results to GIS as an input for further analysis, the only practical issue is data availability or time constraint for collecting new data for the accuracy of the model.

The following section presents the visualisation results for the assigned indicators in the Parramatta study area. The GIS model illustrates GI sustainability performance for current urban settings in three categories (ecological, health and sociocultural). The economic category, which demonstrates the monetary value of number of GI performance does not map and is not visualised through GIS because this category only monetarises four selected indicators from the other categories.

7.5 Review of results Microclimatic modifications can be effectively achieved through changing surface characteristics, urban geometry and by implementing appropriate GI types in the appropriate location.

Effective surface albedo According to the land surface classifications of the study area shown in Figure 7.3, 22.14% of the area is covered with roads and footpaths (mostly asphalt and concrete pavements) and 12.83% comprises dark roofs. Results demonstrate that the overall effective surface albedo score for the area is 2.65 ~ 3.0, which is considered to be a medium level of sustainability performance (Table 7.6). Therefore, to improve this albedo value, replacing less reflective, dark and hard surfaces with green or light-coloured surfaces (e.g. shrubs, grass, white roofs etc.) that have higher surface albedo value is recommended. (For more details on surface types and associated albedo values in an urban area, refer to Table 6.3 in Chapter 6.)

This strategy is also addressed by increasing mass planting areas when a vacant area is available, as well as by substituting conventional concrete and asphalt materials with cool 194 paving (porous and light-coloured pavements) to reduce heat absorption and radiation. Permeable (or porous) pavements have higher solar reflectivity and less surface temperature compared to dark, hard and impermeable surfaces. Also, porous pavements provide the conditions for evaporative cooling, in addition to runoff reduction (flood control) and water infiltration (pollution control). It should be taken into account that permeable pavements cannot be used everywhere (for example, a site where the soil is contaminated or infiltration is not possible).

Table 7. 6 Case study - Effective surface albedo

C1: Climatic and microclimatic modifications C1.1: Effective surface albedo

ECOLO

GICAL INDICATORS

Evapotranspiration rate The overall evapotranspiration performance score for existing GI in the case study is 3.05 ~ 3.0, which is considered to be a medium level of sustainability performance (Table 7.7). As was mentioned earlier, three adjustment factors (ks, kd and kmc) are correlated to the evapotranspiration rate. Therefore, to improve this score in each grid cell, site specific strategies are required as well as consultation with qualified arborists. For example, in the Parramatta area, a number of species that are recommended by arborists are: Platanus orientalis, Lophostemon confertus, Eucalyptus sp., Flindersia australis, Callistemon salignus, Lagerstroemia indica, Callistemon viminalis, Tristaniopsis laurina and Jacaranda mimosifolia.

Most of these species require very low to moderate water use (ks). Some examples are Callistemon viminalis with a low water use rate (0.25), Eucalyptus sp. (e.g. Sugar Gum) - very

195 low (0.1), and Jacaranda mimosifolia - moderate (0.6). In addition to urban tree canopies, examples of recommended native shrubs and their low water requirements include Banksia ericifolia (0.25) and Callistemon citrinus (0.25). Recommended native ground covers are: Grevillea and Yellow buttons (0.1. As a guide, the Green Star rating tool lists the ks for a number of common plant species that can be found in urban areas in Australia to help planners find the appropriate plants and their respective water requirement factors (Green Building Council of Australia 2014).

In addition to the ks factor, species distribution in a landscape zone can impact evapotranspiration rate. However, to achieve a higher density factor value (kd) in each grid cell, which is the goal, it is recommended to plant a variety of types of trees, shrubs and groundcovers to score the ideal ks of 1.1–1.3. Ks will achieve only an average value if only one predominant vegetation type is present. A very high microclimate factor (kmc), which is ideal, will be achieved through planting vegetation in zones which are exposed to high winds and/or sun (e.g. planting near street medians, in wide canyons and parking lots) or exposed to reflected surfaces with high surface albedo, which increases kmc as well as the evapotranspiration rate.

Analysis of the evapotranspiration rate (ETL) in the study area - not surprisingly - demonstrated that commercial zones with high-rise buildings and deep canyons achieve the lowest and least desirable rate. However, the grid cells with low ETL value can be improved by planting a mixture of native trees, shrubs and groundcovers in front of reflective surfaces (a maximum of 70% plant density is effective and efficient in improving ET, as explained earlier). It is worth mentioning that planting in shallow canyons with E-W orientation would effectively change the evapotranspiration rate positively. However, it is worth mentioning that higher ET requires more irrigation water use. In this case perhaps greywater could be recovered from the buildings in the commercial zones, treated and used for irrigation.

196

Table 7. 7 Case study - Evapotranspiration rate

C1: Climatic and microclimatic modifications C1.2: Evapotranspiration rate

ECOLO

GICAL INDICATORS

Ventilation In the case study of Parramatta, GIS and cadastral data were used to calculate aspect ratio and the results were then classified based on street orientations. Results revealed that the aspect ratio (H/W) of selected streets varies between 0.2 and 4.5 with the average value of 3.11. Prevailing winds in coastal Sydney come from the North-East, South, and West. Winds from the North-East tend to be summer sea breezes and bring welcome relief on summer days. Winds from the South occur throughout the year and tend to be cold and associated with frontal systems that can last for several days. Winds from the West tend to be the strongest of the year and are associated with large weather patterns and thunderstorm activity; they occur throughout the year and can be cold or warm depending on inland conditions, see Table 6.11 and Figure 6.3. In this pilot study, the overall performance score of existing GI is 2.60 ~ 3.0 for the ventilation indicator, which is considered to be a medium level of sustainability performance (Table 7.8). Visualisation results show that areas with fewer trees and wide canyons achieve low to medium scores. However, it is recommended that grid cells with low ventilation rates can be improved through modifying the aspect ratio of the canyon (especially for shallow canyons) by planting trees on the appropriate side of the street.

197

Table 7. 8 Case study – Ventilation

C1: Climatic and microclimatic modifications C1.3: Ventilation

ECOLO

GICAL INDICATORS

Shading effect The overall urban GI shading effect score in the study area is equal to 2.76 ~ 3.0, which is classified as a medium level of sustainability performance. Calculation of this indicator is linked to the percentage of leaf area in each grid cell (see Table 7.9). Therefore, to achieve the highest possible leaf area per grid (with a size of 100 m by 100 m), available growing space is filled preferentially with large- to medium-sized trees.

198

Table 7. 9 Case study - Shading effect

C1: Climatic and microclimatic modifications C1.4: Shading effect

ECOLO

GICAL INDICATORS

Leaf area calculation results with iTree Eco demonstrate that, among all species in the study area, Brushbox (Evergreen, Native), Sweet Gum (Deciduous, Exotic), Water Gum (Evergreen, Native), Jacaranda (Deciduous, Exotic), and Bottlebrush (Evergreen, Native) provide the highest leaf areas (m2) in sequence. Therefore, to improve shading effects in grid cells with low LA values, planting trees from the above list, keeping soil type and condition in mind, is recommended (Table 7.10).

Table 7. 10 Case study results - Proposed trees with good shade effect performance (additional information is adopted from Mather and Morton (2008).

Drainage Origin Soil Conditions Requirements

(E) Evergreen / Deciduous (D) Native/Exotic Shale Sandstone Transitional Alluvium Good Average Poor

Botanical Common

Name Name

E (Large tree Lophostemon Brushbox height: 16-20m N - - confertus × × × × × and spread 16m) Liquidambar D (Large tree Sweet straciflua height: 16-20m E - Gum × × × × × × ‘Parasol’ and spread 16m) E (Medium tree Tristaniopsis Water height: 10-12 m N laurina Gum and spread 8-9 × × × × × × × m) 199

D (Medium tree Jacaranda height: 10-12 m Jacaranda E - - mimosifolia and spread 8-9 × × × × × m) E (Small tree, Weeping Callistemon height: 6-8m Bottlebrus N viminalis and spread 4- × × × × × × × h 5m)

Air pollutant removal Air pollutant removal with existing urban GI in the study area ranks as 2.55 ~ 3.0 out of 5.0 which is considered to be a medium level of sustainability performance (Table 7.11). iTree Eco analysis results demonstrate that existing GI air pollution removal is significant for ozone and nitrogen dioxide in comparison with other air contaminants (Figure 7.12). It is estimated that trees in Parramatta CBD remove 198.54 kilograms of ozone (O3) air pollution, 12.57 kg of carbon monoxide (CO), 50 kg of nitrogen dioxide (NO2), 2.59 kg of particulate matter less than 2.5 microns in size (PM2.5), and 12.07 kg of sulphur dioxide (SO2) per year.

Figure 7. 12 Monthly pollution removal by urban GI in the case study Parramatta CBD

200

Table 7. 11 Case study - Air pollutant removal

C2: Air quality improvement C2.1: Air pollutant removal

ECOLO

GICAL INDICATORS

Carbon storage and sequestration The amount of carbon sequestered annually increases with the size and health of the trees. The estimated annual carbon sequestration rate with the existing urban GI in the study area ranks 2.54 ~ 3.0, which is considered to be a medium level of sustainability performance (Table 7.12). To improve the performance of GI and increase the uptake of carbon in each grid cell, it is recommended to increase tree density in each grid while selecting species with a high carbon sequestration rate.

201

Table 7. 12 Case study - Carbon storage and sequestration

C3: Carbon offsets C3.1: Carbon storage & sequestration

ECOLO

GICAL INDICATORS

iTree results also show that gross carbon sequestration of the study area is about 31,185 kilograms of carbon per year, while 494,200 kg of carbon is stored in tree trunks. Of all the species, Lophostemon stores and sequesters the most carbon (approximately 29.4% of the total carbon stored and 37.8% of all carbon sequestered.) (refer to Figure 7.13).

Figure 7. 13 Annual gross carbon sequestration with existing urban GI in the case study area Parramatta CBD

202

Energy saving Trees affect energy consumption by shading buildings, providing evaporative cooling, and blocking winter winds. Trees tend to reduce building energy consumption in the summer months and can either increase or decrease building energy use in the winter months, depending on the location of the trees around the building. Estimates of tree effects on building energy use are based on field measurements of tree distance and direction to buildings. A summary of cooling and heating energy saving results in the Parramatta case study is shown in Table 7.13. The results demonstrate trees in the study area do not provide any heating benefits. Also, they may have negative impacts by increasing building energy use during winter time due to having been planted in locations which shade the buildings during winter. But this is very minor and the GIS modelling only considers the cooling benefits of urban trees.

Table 7. 13 Case study, annual energy savings due to trees near buildings Heating Cooling Total MBTU -11 n/a -11 MWH -1 106 106 Carbon avoided (tonne) 0 30 30 MBTU=one million British Thermal Units MWH= Megawatt-hour Note: Negative numbers indicate that there was not a reduction in heating, cooling and/or carbon emissions value, and rather that carbon emissions and values increased by the amount indicated as a negative value.

The overall energy saving score for cooling in the study area is equal to 2.54 ~ 3.0, that is classified as a medium level of sustainability performance (Table 7.14). Therefore, to achieve the highest possible leaf area per grid (100 m by 100 m), available growing space should ideally be filled with large- to medium-sized deciduous trees (deciduous, to minimise the need for additional winter heating).

203

Table 7. 14 Case study - Energy saving

C4: Reduced building energy used for cooling and heating C4.1: Energy saving

ECOLO

GICAL INDICATORS

Water quality To calculate the water quality indicator, the three most common water contaminants were selected to include in the model: total nitrogen (TN), total phosphorus (TP) and total suspended solids (TSS). Average annual rainfall data in Parramatta CBD has been obtained from the North Parramatta weather station (0066124), which is 973.5 millimeter per year averaged over 1965-2016 (Australian Bureau of Statistics 2017). Analysis of the results shows: - TSS pollutant load ranges between 55.09 kg/ha yr and 1233.69 kg/ha yr - TP pollutant load ranges between 0.09 kg/ha yr and 2.28 kg/ha yr - TN pollutant load ranges between 0.70 kg/ha yr and 14.07 kg/ha yr Existing GI types in the study area consist of tree canopies that are located over grass (e.g. parks) or street trees planted over impervious surfaces. Table 6.18 explained the contribution of these two types of GI to remove pollutants is less than 23.8% and this discounting rate does not meet the national target that are 80%, 60% and 45% reduction for annual loads of TSS, TP and TN, respectively. As this study does not test any design scenarios (e.g. adding series of BMPs), baseline values ranging between 1 and 5 are applied on each grid cell, based on the pollutant load (sum of three contaminants) by considering the amount of pollutants removed annually by existing GI (street trees in this case). The overall water quality score due to impacts of existing GI in the study area is equal to 2.98 ~ 3.0, which is classified as a medium level of sustainability performance (Table 7.15). Therefore, to achieve the national targets, installation series of BMPs in the ascending order of pollutant uptake rates is recommended. 204

Table 7. 15 Case study - Water quality

C5: Hydrological regulation C5.1: Water quality

ECOLO

GICAL

INDICATORS

Avoid surface runoff Trees and shrubs intercept precipitation, while their root systems promote infiltration and storage in the soil. To calculate the runoff reduction indicator in Parramatta CBD, iTree Eco was utilised. Results revealed that existing GI helps to reduce runoff by an estimated 403 cubic metres a year with an associated value of A$910. Figure 7.14 shows 10 species that provide greatest impact on overall runoff reduction and their associated avoided stormwater cost.

205

Figure 7. 14 Avoid runoff and value for species with greatest overall impact on runoff

The overall avoid runoff score for existing GI in the study area is equal to 2.60 ~ 3.0, which is classified as a medium level of sustainability performance (Table 7.16). Therefore, to achieve the greatest runoff reduction outcome, planting more species in watershed that are recommended in Figure 7.14 will improve the overall performance effectively and efficiently.

Table 7. 16 Case study - Avoid surface runoff

C5: Hydrological regulation

C5.2: Avoid surface runoff ECOLO

GICAL INDICATORS

206

Promoting conservation (Habitat connectivity)

230 native fauna species have been recorded in the Parramatta LGA (Parramatta City Council 2015). A total of 32 fauna species have been listed as threatened and vulnerable. The list contains 16 birds, 12 mammals, two frogs, one reptile and one invertebrate.

To calculate habitat connectivity the least-cost model was applied. This model requires two sets of input data; a resistance map; and the recorded location of the specific species. Five factors have been included in calculating the resistance score: land-use, land cover, slope, distance to water bodies (Parramatta river and streams) and distance to main roads and highway. To calculate slope, 10 metre DEM data has been used.

Table 7.17 shows results from mapping five factors or sub-indicators to obtain resistance score and habitat connectivity of the study area by overlapping these five maps. The overall connectivity score is equal to 2.46 ~ 2.0, which is classified as a low-medium level of sustainability performance. Not surprisingly, this indicates the level of habitat connectivity is medium-low in the dense urban CBD. This level of habitat discontiuity is more highlighted in the centre and south of the Parramatta CBD. To make connection between isolated patches, it is recommended to plant locally native trees such as the Cheese Tree in public and private lands to promote biodiversity and habitat linkages.

Table 7. 17 Case study – Habitat connectivity

C6: Biodiversity protection and enhancement C6.1: Promoting conservation (Habitat connectivity) Fauna Land-use Land cover Slope Distance to Distance to

ECOL highways water INDICATORS

GICAL

207

Note: red dots show the location of the fauna species were provided by the Parramatta City Council between 2011 and 2015

Species diversity

According to the Parramatta biodiversity strategy 2015-2025, the Local Government Area of Parramatta includes “17 distinct ecological communities. Twelve of these native vegetation communities are listed under the NSW Threatened Species Conservation Act 1995 (TSC Act) as endangered or critically endangered ecological communities (EEC). Four of these eleven EECs are also listed in Parramatta LGA under the Commonwealth Environment Protection and Biodiversity Conservation Act 1999 (EPBC Act)” (Parramatta City Council 2015). More than 600 native flora species have been recorded in Parramatta LGA. A total of 13 of them are identified to be threatened and vulnerable species including Acacia pubescens, Callistemon linearifolius, Epacris purpurascens var., Grevillea parviflora subsp., Leptospermum deanei, Persoonia nutans, Pterostylis saxicola, Zieria involucrata Wilsonia backhousei, Wahlenbergia multicaulis, Tetratheca glandulosa and Pomaderris prunifolia.

In the study area, there are 1780 street trees with species richness of 53. To calculate diversity score or Shannon’s equitability, the Shannon diversity index was applied for whole given area then divided by the maximum diversity (lnS), where S is total number of species in the community (richness). Results indicate the Shannon diversity is equal to 2.9 for whole site with the equitability score of the species diversity is 0.73 (=2.9/ln (53)). It means the site is 73% diverse when compared with the maximum possible diversity for the study area. In the other words, it is a medium-high diverse community. Refer to Appendix C for the detailed information to calculate species diversity in the whole study area.

208

The same procedure was applied to calculate species diversity in each grid cell. It is important to consider species on private lands in future studies. However, in this pilot study of Parramatta a species inventory on private land (backyards, roof gardens etc) was not available. Therefore, only street trees have been considered (Table 7.18).

Table 7. 18 Case study – Species diversity

C6: Biodiversity protection and enhancement C6.2: Species diversity (Flora)

ECOL

GICAL INDICATORS

Improving physical well-being (Physical accessibility)

300 metres proximity walking distance to urban green spaces was used as corresponding to approximately a seven minutes’ walk along walkable pathways. The calculation was performed through GIS by Network Analysis and Spatial Analyst extension toolboxes. In three steps: - Extract footpaths in the study area and run Euclidean Distance tool to create a surface in which each raster cell has a value of distance between the closest footpath and itself (Figure 7.15). - Then create a travel cost surface between each raster cell and edge of nearest park along the closest path that was created in the first step (Figure 7.16).

209

- Normalise distance values in metres by reclassifying ranges between 1 and 5 and cut off point of 300m (Table 7.19).

Figure 7. 15 Euclidean distance - footpath analysis Figure 7. 16 Travel cost surface analysis

The overall physical accessibility score within 300 metre walkable distance to urban greenery in the study area is equal to 3.35~ 3.0, which is classified as a medium level of sustainability performance (Table 7.19).

Table 7. 19 Case study – Physical accessibility

C7: Improving physical well-being

S R O T A C I D N I H T L A E H

C7.1: Physical accessibility (Proximity in walkable distance)-300m

210

Improving social well-being (Equity: diversity in the spatial distribution of GI services)

There are 304 green spaces in Parramatta LGA. 37 of them are located within a 500 metre buffer zone around the study area. To calculate the equity score or the diversity in spatial distribution of GI services, the Shannon diversity index was applied for the whole study area as well as for individual grid cells. Results indicate the Shannon diversity is equal to 1.73 for whole site. To understand how diverse is this result, we divided this score by maximum diversity of GI services in the study area which is natural logarithm of number of GI classes (refer to Table 7.20) which is seven classes in this pilot study (ln7). Therefore, the equity score of the functional diversity is 0.89 (=1.73/ln (7)). It means the site is 89% diverse in terms of GI equity to compared with the maximum possible functional or services diversity of the study area. This score indicates that the study area has a medium-high equity level of services distribution by GI types. This score is consistent with the average result obtain from applying Shannon diversity index on individual grid cells.

Table 7.20 classifies these 37 GI in seven groups based on their size, facilities and functional activities. This table also shows the service radius for each individual GI class that was considered in the calculations.

Table 7. 20 Functional classification of GI types and their service radius zone Type (i) Number (ni) Service Radius Mini-park/pocket park 12 300m Neighborhood park 7 500m Community park 3 800m 211

Large urban Park 1 2km Sport park 2 1km Linear park-trails 6 300m Nature preserve 6 300m

Figure 7.17 and 7.18 illustrate the functional distribution of GI types and service areas covered by individual types respectively. These illustrations indicate grid cells that are located within each service zone. Almost all grids at least have access to one service zone.

Figure 7. 17 Spatial functional distribution of GI types

Figure 7. 18 Service radius covered by individual GI types The overall GI equity score (as a proxy indicator for social well-being) is equal to 3.66 ~ 4.0, which is classified as a medium-high level of sustainability performance (Table 7.21). Therefore, to achieve the highest possible score per grid (100 m by 100 m), increasing the spatial distribution of various GI types in terms of their functional diversity is recommended. 212

Parramatta Park is the largest green patch in the study area that provides diverse functions and services for at least two kilometres’ buffer zones. However, mapping GI equity (Table 7.21) shows lack of distribution of other GI types near Parramatta Park.

Table 7. 21 Case study – GI equity

C8: Improving social well-being C8.1: Equity (Diversity in the spatial distribution of major GI types)

HEALTH HEALTH

INDICATORS

Improving mental well-being

The overall visual accessibility score in the study area of parramatta is 2.90 ~ 3.0, which is classified as a medium level of sustainability performance. Figure 7.19 illustrates visible buildings’ façade analysis based on the area of selected façades that have views to the greenery (length multiplying by building height). Note that tall buildings have more opportunity for greenery exposure.

213

Figure 7. 19 Visible building’s facade analysis

Table 7.22 indicates that the central part of the study area which has taller buildings, achieved a higher score. This indicator considers the view from buildings to open green spaces. Therefore, grid cells that include building blocks are only considered in this calculation.

Table 7. 22 Case study – visual accessibility

C9: Improving mental well-being C9.1: Visual accessibility (view-based exposure)

HEALTH HEALTH

INDICATORS

214

Food production- domestic gardens

The overall land productivity capacity performance score (community gardens and home gardens) for producing domestic food is 3.04 ~ 3.0, which is considered to be a medium level of sustainability (Table 7.23). For this calculation, population data taken from mesh blocks were converted to the grid cell and then rear gardens were calculated based on lot size (50 percent of the total required private open space in each lot), as explained in Chapter 6. This calculation only considered residential zones (backyards). Therefore, the visualisation results show commercial and industrial zones with no data. In practice, there would be opportunities for development of community gardens in non-residential land-use zones as well.

Table 7. 23 Case study – Food production

C10: Food production C10.1: Land productivity capacity – community gardens and rear gardens (residential zones)

SOCIOCULTURAL SOCIOCULTURAL

INDICATORS

215

Mixed land-use neighbourhood

Mixing GI with different land uses increases recreational and ecotourism opportunities. The overall entropy score (as a proxy sub-indicator for mixed land use) is 2.59 ~ 3.0, which is classified as a medium level of sustainability performance (Table 7.24). To achieve the highest possible score per grid (100 m by 100 m), increasing spatial distribution of land use and retrofitting with various GI types in terms of composition and diversity is recommended.

Table 7. 24 Case study – Mixed-use neighbourhood

C11: Opportunity for recreation and ecotourism C11.1: Mixed-use neighbourhood (mixing GI with different land uses)

SOCIOCULTURAL SOCIOCULTURAL

INDICATORS

216

Walkability design score

The walkability design score can assist in assessing pathways in accordance with GI planning standards, (assigning scores based on design criteria). Figure 7.20 shows existing footpaths and cycleways in the Parramatta CBD study area.

Figure 7. 20 Footpath and cycleways

The overall walkability score in the Parramatta study area is 3.24 ~ 3.0, which is classified as a medium level of sustainability (Table 7.25). To achieve the highest possible score per grid (100 m by 100 m), improving footpath quality by planting trees to provide shade and adding green buffer zones between cycleways and driveways/roads for cyclists’ safety is recommended.

217

Table 7. 25 Case study – Walkability

C12: Improving pedestrian ways C12.1: Walkability

SOCIOCULTURAL SOCIOCULTURAL

INDICATORS

Connectivity (intersection density)

Figure 7.21 illustrates parks as interest points to attract visitors and the normalisation value of network analysis in the study area by considering intersection density value.

Figure 7. 21 Network analysis and interest points (green open spaces).

218

The overall connectivity score in the Parramatta study area is 3.59 ~ 4.0, which is classified as a medium-high level of sustainability performance. Table 7.26 indicates connectivity value near interest points (green open spaces). Grid cells with a higher value have a higher number of intersections and fewer dead-ends, while being well connected to neighbouring nodes with high connectivity value.

Table 7. 26 Case study – Connectivity (intersection density)

C12: Improving pedestrian ways C12.2: Connectivity (intersection density)

SOCIOCULTURAL SOCIOCULTURAL

INDICATORS

219

7.6 Overall results

Table 7.27 shows the standard deviations and overall results for indicators in three categories: ecology, health and sociocultural. The overall results for the Parramatta CBD study area within these three categories amongst 18 sub-indicators is 2.89 ~ 3.0, which is equal to a medium level of sustainability performance. By tracking back to the results assessing individual indicators (Table 7.27), it is revealed that ecological performance of existing GI (~ 2.69) is lower than the overall score (~ 2.89) while it is higher in health category (~ 3.30) and sociocultural (~ 3.12).

Table 7. 27 Overall score of existing GI sustainability performance in Parramatta CBD

Overall Standard CATEGORY INDICATOR SUB-INDICATOR score Deviation (mean) C1-1: 1.41 2.65 Effective surface albedo C1-2: 1.41 3.01 Climatic and microclimatic Evapotranspiration rate C1 modifications C1-3: 1.15 2.61 Ventilation C1-4:

Urban

canopy

geometry 1.47 2.76 Shading effect C2-1: C2 Air quality improvement 1.25 2.56 Air pollutants removal C3-1: C3 Carbon offsets 1.38 2.54 ECOLOGICAL Carbon storage & INDICATORS sequestration Reduced building energy C4-1: C4 1.11 2.54 used for cooling and heating Energy saving C5-1: 1.43 2.98 Water quality C5 Hydrological regulation C5-2: 1.22 2.60 Avoided surface runoff C6-1: Promoting conservation 1.05 2.46 Biodiversity-protection and (Habitat connectivity) C6 enhancement C6-2: Enhancing diversity of 1.35 2.91 species (Species diversity) C7-1: Improving physical well- Physical accessibility C7 1.20 3.35 being (proximity in walkable distance) Improving social well-being C8-1: HEALTH (e.g. social interaction; Equity INDICATORS C8 1.10 3.66 social integration; (diversity in the spatial community cohesion) distribution of GI types) Improving mental well-being C9-1: C9 (e.g. reduced depression and Visual accessibility 1.17 2.90 anxiety; recovery from (view-based exposure)

220

stress; attention restoration; positive emotions) C10-1: Land productivity capacity (community C10 Food production gardens and domestic rear 1.42 3.04 gardens)

SOCIOCULTURAL Opportunity for recreation C11-1: Land use mixed INDICATORS C11 1.34 2.59 and ecotourism Improving pedestrian ways C12-1: Walkability 1.13 3.24 (Safety, quality of path and C12 C12-2: Connectivity linkage with other travel 1.30 3.59 modes)

The general notion to improve ecological performance of GI is to rethink which existing plant species could be replaced by more appropriate native species.

In addition, the case study model results suggest GI improvement strategies that can be incorporated into urban planning schemes in order to achieve low carbon output, environmental sustainability and social and eco-friendly cities. These recommendations are summarised as follows.

1. Use materials with high albedo value (e.g. replace dark roofs with cool and green roofs). 2. Plant native species with low to moderate water required (ks) as well as develop the strategy to distribute planting zones with a mixture of dense trees, shrubs and ground cover to achieve a higher value density coefficient (kd). A very high microclimate factor (kmc) will be achieved through planting greenery in zones which are exposed to very high winds and/or sun (e.g. planting near street medians and in wide canyons, especially those with E_W orientation, and parking lots) or close to reflective surfaces with high albedo (e.g. buildings with a reflective facade) that will increase the evapotranspiration rate. 3. Modify the aspect ratio of the canyon by planting trees on the appropriate side of the street. Where there are budgetary constraints, preferentially, it is more effective to prioritise and plant first in wide canyons with low aspect ratio (H/W). 4. Increase the percentage of leaf area (LA) in each grid as this will increase the shade effect for pedestrians' comfort and moderate indoor temperatures for adjacent buildings. Fill available growing space with large trees first (16-20m height with a 16m crown spread), then medium-sized trees (10-12m height with a 8m crown spread), and finally small trees (6-8m height with a 5m crown spread). This planting strategy may not be practical or desirable in all situations, but the goal of improving the shading effect performance of GI in the assessment model is linked to maximising the leaf area in each grid cell. Generally, spacing between tress should be: 5-7m centres for small trees; 7-10m centres for medium trees; and for large trees at 10-15m centres. 5. To increase GI performance in the uptake of air pollutants, select the right plant species and plant in the right location. Coniferous tress (for example, hoop pines, Callitris, Wollemi pines) are more effective in trapping pollutant particles than deciduous trees. In addition, selecting species which are low level BVOC-emitters can lower overall BVOC emissions

221

and thereby decrease ozone levels in urban areas. Analysis with iTree Eco indicates that urban trees in the study area emitted an estimated 390 kilograms per year of volatile organic compounds (VOCs) (340 kg of isoprene and 60 kg of monoterpenes). However, emissions vary among species (e.g. some genera, such as oaks, are high isoprene emitters) and the amount of leaf biomass. Fifty-two percent of the urban tree canopies’ VOC emissions were from Eucalyptus and Lophostemon species These VOCs are precursor chemicals to ozone formation. 6. Maximise carbon sequestration performance in grid cells with low sustainability rate, increase tree density in each grid while planting species with a higher rate of sequestration. According to the iTree analysis results, seven species that provide higher GI value in sequestering carbon dioxide in the study area are: Oak spp. (e.g. Quercus robur), Blue Gum (Eucalyptus saligna), Fig (Ficus macrocarpa), Paperbark (Melaleuca quinquenervia), Planetree (Platanus acerifolia), Brushbox (Lophostemon confertus), and Jacaranda. 7. To achieve higher energy saving benefits, in particular for cooling, planners in the southern hemisphere should plant native coniferous or evergreen tree species on the east and west sides of a building and plant deciduous trees on the south and north sides. According to the iTree analysis results from the case study, Spotted Gum (Corymbia citriodora), Australian Teak (Flindersia australis) and Brushbox (Lophostemon confertus), which are large evergreen trees, provide the highest cooling value in the pilot study when they were planted in the ideal location. This is also true for Platanus orientalis, which is a large deciduous tree. 8. Surface water quality indicator can be improved through installation of a series of technical GI types or BMPs. Each BMPs has different capacity to remove water pollution, so selection of the appropriate type of BMPs depends on the site’s pollution loads and dominant pollutant. It is recommended in a watershed-scale, the BMPs are installed in the ascending order of their capacity to remove pollutant. 9. Runoff reduction is influenced by the soil type, gradient, permeability and land use/land cover. These factors are all interpreted and summarised in the runoff coefficient (Rvu). To control runoff effectively, designers must select material and finishes for land cover with appropriate (Rvu) value by considering gradient and land use. In addition, various BMPs such as infiltration devices, bioretention and permeable pavement have high capability to reduce the overall volume of runoff. Analysis of the results shows Lophostemon and Platanus provide greatest impact on overall runoff reduction among other species.

Economic value of the study area

Survey participants selected four indicators out of ten from the initial set of economic indicators: the value of avoiding CO2 emissions and carbon sequestration, avoiding energy consumption through cooling and heating, air pollution removal, and reducing the cost of using private cars by encouraging people to walk and cycle. Three of these selected indicators are derived from the ecological category calculated by iTree Eco. Therefore, the monetary value that was used in the tool is applied here too. The value of the fourth indicator was selected through the online survey to monetarise based on the value for changing travel mode from 222 passive to active in accordance with the provision of green and high quality pathways and cycle ways (per kilometre). To do this, the total length of the green pathways in kilometres is estimated then multiplied by $121.91. In this study area, total footpath length is 53.940 km, of which 37.86 km is greened and thus the monetary value of the fourth indicator is A$4,615.5 (per year).

Table 7.28 indicates dollar value of the selected indicators in economic category in study area of Parramatta CBD. Table 7. 28 Case study – Dollar value of selected economic indicators

C13 Value of avoided CO2 and carbon sequestration A$711 per year C14 Value of avoided energy consumption A$4,600 per year ECONOMIC C15 Value of air pollutant removal/avoidance A$222 per year INDICATORS C16 Reducing cost of using private car by encouraging A$4,615.5 year people to walk and cycle in green network

The overall economic benefits of four selected indicators seems to be very low. But it is worth mentioning that the indicators that are screened from this category may provide more benefits to health and well-being.

Summary

This chapter has applied the proposed indicator-based model to the selected study area in Parramatta to calibrate the model and assess the multidimensional performance of urban GI. To do this, the study area was divided into a 100×100 m grid via ArcGIS software, and each indicator calculated in each grid cell individually applied based on the proposed equations. Since the different indicators used different measurements, scales and units, a normalisation technique was utilised to convert the indicator value as expressed in various units into the five, relative, normalised baseline values (ranging from 1 to 5 - low to high level of sustainability). Then, results from all 16 indicators and sub-indicators (in three categories) were aggregated to demonstrate the overall performance of GI in each category. It is noted that, during the model testing, all indicators were assumed to have the same level of importance and the weighted value for all of them considered to be equal to 1. This is because this was a test process for calibrating the model and keeping all indicators at the same level of importance helps to identify errors irrespective of external variables.

The weighted value for each indicator is based on the experts’ survey results (Chapter 5 and Appendix B). In the future, when applying the model to another study area, a weighted value can be assigned to each indicator based on the proposed weight values in Chapter 5, or they can be set up by decision makers based on the objective of the project (if required). The overall GI sustainability performance index (SPI) can then be calculated as follows:

SPI w x Eq. 7. 1  ii

Where, wi and xi are the weight and normalised value for the indicator, respectively. 223

CHAPTER 8: CONCLUSIONS

Introduction

Assessing and monitoring sustainability levels for natural and built environments has become crucial in recent years because of global challenges in coping with accelerating urbanisation as well as human-driven land alteration, transformation and fragmentation, and of course climate change. Green infrastructure (GI) has been identified as an alternative, nature-based and cost- effective sustainability solution for promoting more sustainable urban development and restoring degraded ecosystems. A number of sustainability-oriented frameworks have been developed at the project and policy level to assess social, economic and ecological dimensions of sustainability. The main challenge in these sustainability assessment frameworks is identifying an appropriate and comprehensive, set of scientific, practical and policy-based indicators which help to effectively monitor sustainability progress in the effort to build a sustainable environment and society.

The main purpose of this research has been to develop an indicator-based assessment model as a means to evaluate the performance of urban green infrastructure that is comprehensive, integrated and multi-scale. Many local government agencies in Australia and elsewhere have been prioritising increasing tree canopy as part of their urban forest strategy in order to mitigate high surface and ambient temperatures and urban heat island effects. However, implementing tree canopies and other types of GI have limitations in an urban area compared with open spaces or suburban areas because of constraints in available space, street width, building height, street orientation and surface characteristics. These factors also influence the capability of GI to deliver benefits, particularly in microclimatic performance. Thus, it is vital to understand under which circumstances these projected amounts of greenery should be configured and distributed to reach the desired sustainability level and thereby to provide the greatest benefits to pedestrians and urban residents while sustaining ecosystem functions with limited amount of greenery.

Successful GI planning requires understanding its multiple benefits and functions through an inclusive way of measuring these benefits. There are several methods and models to evaluate the benefits of urban ecosystems and landscape performance, as discussed in Chapter 2 but it remains unclear how to bring natural science-based landscape functions together with more comprehensive estimates of all services delivered by ecosystems. Therefore, the primary research objective has been the development of a single, integrated and comprehensive model to evaluate multiple aspects of GI key benefits and their contribution to reaching sustainability targets in projects while reducing uncertainties.

224

8.1 Discussion of findings in relation to the research objectives

An extensive review of the existing literature, frameworks and assessment models within green infrastructure research studies found that there are only a few tools available that assess the performance of green infrastructure, let alone a single integrated tool that can be applied across a wide range of criteria. Moreover, the review in Chapter 2 has shown that there is a lack of consistency among the tools available (Chapter 2). This also inhibits the ability to compare costs and the associated benefits of green versus grey infrastructure and in turn makes it difficult to convince decision makers to invest in GI.

In order to determine the most effective criteria for such an approach, four major research objectives were formulated:

1. Establishment of a conceptual framework;

2. Investigation of key indicators;

3. Development of a measurement system for indicator-based model; and

4. Validation of the proposed model via a case study.

The model as developed allows the investigation of the impacts of urban development on existing GI, testing various scenarios of future development, and, based on established sustainability targets, identifying the areas that require new GI implementation or improvements. The schematic diagram in Figure 8.1 demonstrates the model development process.

Review and refine the Develop initial list of Review, weight, screen proposed conceptual indicators based on Establish a conceptual and aggrigate framework through DPSIR framework and framework (DPSIR) indicators through semi-structured interviewees' attitudes online questionnaire interviews (thematic concept)

Validate and verify Test indicator-based Formulate each Develop key indicators, baseline model in a pilot study indicator and establish performance indicators value and equations area baseline value

Figure 8. 1 Overview of development of the model

225

8.1.1 Establishment of the conceptual framework The first research objective was to develop a conceptual framework as a foundation for the model based on a literature review and adjusting it for the Australian context. Based on this objective, the first research question was:

What is the research philosophy (conceptual framework) when identifying GI performance indicators?

The literature review has shown that the key theoretical fundament behind green infrastructure is the ethical relationship between humans and nature, also known as environmental ethics. Environmental ethics is rooted in the early writings of John Muir, Albert Schweitzer, Aldo Leopold and Roderick Nash. It is argued that the conflict between ecocentrism and anthropocentrism in society is recognised as one of the most significant ecological moral dilemmas and this conflict also explains the prevalence of paradoxical environmental ethics decisions. To come to decisions that equally consider humans’ and nature’s rights, planners and decision-makers must consider the possible consequences of a variety of different options and determine which consequence should take priority. This research study has the same fundamental approach in mind: the ideal balance between unavoidable human activity and its environmental impacts on the one hand, and implementing the green infrastructure concept into the urban planning process to minimise these negative impacts on the other hand (dual rights-based approach). The DPSIR (Driving force–Pressure– State–Impact–Response) model, which has been widely used by the European Environment Agency (EEA) to develop environmental management and landscape assessment frameworks was utilised to demonstrate how different aspects and dimensions of GI planning contribute to mitigating the negative impacts of, for example, urbanization and human activity. This framework, which is the foundation of this research is the initial list of indicators that have been determined (Figure 8.2). It demonstrates the link between human activity and green infrastructure planning as a nature-based solution to reducing negative environmental impacts and enhancing urban resilience.

Figure 8. 2 The DPSIR framework showing the link between human activities and GI performance. 226

This DPSIR framework also plays a key role in establishing the environmental, social, health and economic objectives to be achieved with the GI approach, as well as in creating a shared vision and consensus about GI strategies in response to natural capital depletion. In addition, this framework employs indicators to determine the benefits GI brings by specifically looking at what humans, as well as ecosystems and surrounding natural environment, gain from it. As such, the ecological benefits resulting from GI lead to enhanced human health and well-being while preserving and maintaining the environment and natural resources, which is in line with the dual rights-based approach; redressing the balance between human’s rights and nature’s rights. Therefore, implementing GI in urban settings provides the missing link between humans, nature and the built environment.

8.1.2 Investigation of key indicators

The second objective was to develop a set of indicators in order to define the logical and rational relationships between the various aspects of sustainability in relation to green infrastructure performance and to identify the importance of each indicator.

What are the indicators in use? Are they sufficient to assess and measure the performance of green infrastructure networks comprehensively and in an integrated manner? What are the interactions amongst indicators? What is the order of importance of each indicator? That is, is there a hierarchy amongst indicators in any given assessment ‘system’?

A mixed-method approach (literature review, interview and questionnaire) was used to develop this indicator-based model. The strengths and weaknesses of any assessment model depend on the quality and suitability of the selected indicators so an initial list of indicators was proposed by experts through semi-structured interviews and online questionnaire methods. Semi- structured interviews were conducted with 21 Australian experts who have expertise within the architecture, landscape, urban planning and ecology professions across three sectors of the economy (academic, government and practitioners). The purpose of the interviews was to seek an understanding of the current perceptions and knowledge of GI in the Australian context, and to determine a list of indicators that best evaluate the performance of GI in the built environment. The results revealed nine of the most important GI thematic concepts that could be added into the DPSIR framework. These nine concepts were then classified into three major categories: economic growth; environmental sustainability; and health and well-being (sociocultural is considered a sub-category of health).

Upon an additional review of the GI scientific literature using the assistance of the DPSIR framework, an initial set of 30 GI performance indicators was developed. This initial performance indicator selection received feedback from 373 national and international experts

227 through an online questionnaire that weighed, screened, and aggregated the selected indicators leading to the establishment of an integrated indicator model. The Weighted Average Index (WAI) method was used to analyse indicators. The final selection of indicators was based on a cut-off point of interquartile range (median), where only the indicators above the median (50th percentile or WAI% ≥ 80%) were selected to be a part of the final assessment model.

The model comprises a reduced set of sixteen indicators based on experts’ opinions within four subcategories: ecological, health and well-being, sociocultural and economic - each representing key interactions between human health, ecosystem services and ecosystem health, which is in line with proposed frameworks by other scholars. In terms of the order of importance of each indicator, the results revealed that participants pay greatest attention to human health (25.9%) and ecological aspects (32.4%) of GI performance. For instance, they considered ‘Improving physical well-being’ from the health category and ‘Microclimate modifications’ from the ecology category to be the most important indicators and ‘Reducing the cost of using a private car by increasing walking and cycling’ from the economic category to be the least important.

8.1.3 Development of metrics for indicator-based model The third objective was to determine and define appropriate equations and methods for formulating indicators and establishing the baseline sustainability performance value for each of them. The research questions were as follows:

How effective is each indicator in its actual measurement of sustainability performance? When assessing the degree of sustainability performance of a given green infrastructure, how does it compare to the baseline values?

The selected indicators have different units and scales that cannot be aggregated in the normal way by summing up the original results. Therefore, to normalise them, baseline values ranging between 1 and 5 were used. They represent maximum and minimum values and refer to different contribution levels of sustainable green infrastructure for each indicator. These baseline values have been determined from other applicable studies in Australia. (In the case study, the baseline value was proposed specifically based on the Parramatta study area and can be adjusted in accordance with the project objectives and desired targets set by stakeholders and decision makers.)

Significant points to be made about the model are as follows.

- The model was developed based on the DPSIR framework. This framework helps to clarify the complex relationship between cause and effect variables, and is proven to be the most effective framework in understanding environmental issues and identifying potential GI planning solutions.

228

- The final list of indicators was established based on international survey results with input from key experts, highlighting the potential to use this model worldwide. - This model serves as a rating tool that assists landscape/urban designers and decision makers in identifying challenges will be faced and acting accordingly. - The model provides an overview of the sustainability performance of existing green infrastructure and model output can be useful in setting up policies and sustainability targets to improve sociocultural, health and environmental performance. - This is a multiscale model from micro to macro. It employed iTree Eco for analysis of the environmental benefits of GI and iTree is unique in that it can estimate the ecological benefits of a single tree or shrub and does not have any limitations in terms of scale.

8.1.4 Validation of the proposed model via a case study

The fourth objective was to validate and verify the model by testing it in a pilot case study area. Parramatta CBD, in the Sydney region, was chosen to identify any limitations and strengths of the model. The research question was:

What are the strengths and limitations of the proposed assessment model?

Through testing the model on the performance of existing green infrastructure in a dense urban area, it was found that the application of the 16 proposed indicators was feasible but with some limitations. These limitations were explained in Section 8.3. In terms of the model's strengths, generally speaking, this is the first model that is integrated and comprehensive in measuring GI performance. Also, testing it with a case study supported the model’s capability to provide information referring to the four dimensions of GI sustainability performance - ecology, health and well-being, sociocultural and economic.

8.2 Significance of the research

There is a need for increased awareness of GI at the planning and implementation phases of urban development, both in the policy and practice sectors. An example of this can be found in the ‘Smart Cities Plan’ framework, published by the Australian Government (2016). This framework presents a new vision for smart growth whereby investing in GI along transport hubs and networks has been highlighted as a way to increase walking and cycling and thereby improve environmental outcomes and encourage healthy lifestyles. Other examples of the use of the GI concept can be seen in the GI Technical Guidance Toolkit released by The US Environmental Protection Agency (EPA) (2016), London’s Green Grid (Greater London Authority 2012), Maryland’s GreenPrint, Singapore’s ABC (Active, Beautiful, Clean)

229 guidelines and the European Commission's communication on GI visions, as well as the more recent European Union (EU) project ‘GREEN SURGE’, which is used to implement green infrastructure and urban biodiversity.

One of the main challenges that government agencies face in proposing these policy guidelines is how to actively manage changes in land use and land cover to meet human needs while also protecting and improving ecosystem services and biodiversity. There has been limited research that assesses the multifunctional aspects of urban greenery and its associated benefits exclusively so, this research has established a conceptual framework and identified specific key indicators for measuring the level of sustainability of GI. In addition, formulae and assessment methods, as well as baseline values for each indicator, have been developed. The data visualisation results were demonstrated via ArcGIS software, which makes it possible to develop this model as a user-friendly tool written model-builder in future study. Then, it can be used by local government agencies (e.g. to identify issues with improving or redesigning the sites and setting up tenders, developing design guidelines and strategies); by the planning departments (e.g. land zoning), and by urban planners and landscape designers to test their proposed design scenarios and compare them with the base case to end up with the best option. This study has not only contributed to the existing literature on urban ecology and sustainability through the GI planning approach, but it also has provided a new way of thinking by quantitatively assessing the level of sustainability in an urban setting.

It is hoped that this model will help industry professionals make decisions that will shape the well-being of cities and their inhabitants and determine the value of the benefits of sustainable urban development inclusively. This facilitates sustainability comparisons over time as a means to monitor changes in 16 indicators and to see if the overall results are progressing towards a better or worse level of sustainability. Accordingly, the results generated from the model developed in this thesis are likely to allow evidence-based policy analysis and a framework for governmental agencies and policymakers so as to assist them in keeping each square metre of land as sustainable as possible as a legacy for future generations.

8.3 Limitations and recommendations for future research

This model has been applied to a case study area in Parramatta CBD to validate the model's equations and baseline values, to determine the range of data required and to identify uncertainties and possible errors during generation and overlapping of individual indicators.

It provides fundamental information that helps designers, decision makers and developers to identify issues in various aspects of sustainability in relation to GI as well as guide them to act accordingly to mitigate the negative impacts of future development on existing GI and propose effective GI planning. However, this model, similar to other assessment frameworks, has its weaknesses.

230

It is recommended that future studies simplify the process while including more indicators simply by lowering the cut-off point e.g. from 80% to 70% or less, as well as attempt to propose formulae that require less input data and variables to run the model. In addition, to avoid complexity and miscalculation for such an assessment model, users of the model are advised to choose indicators that are relevant to the object of their project or site’s specific issues. For example, if the local council aims to enhance walkability and connectivity of the neighbourhood through implementing GI, but there are limitations in time, budget and existing data, decision makers and planners could exclusively select indicators from sociocultural and then health categories to use in the model accordingly.

The second limitation of the model is availability and quality of data that are essential to making an assessment accurate. In the Parramatta CBD case study, a complete street tree database for the pilot area was available, but there was neither any data for shrubs and ground cover species nor information about trees in backyards and private lands. Thus, some indicators such as biodiversity, which require data for all species in the study area in the model testing, only considered the street trees because of lack of data. In this case, it is recommended to omit the indicator or run the model based on the best available data and by assigning a weight based on experts' attitudes to that indicator, tolerate the missing data.

The third limitation in the development of the model is the baseline values that are proposed to normalise and classify the outputs from 1 to 5 in order to demonstrate the contribution of GI in sustainability. However, as was mentioned earlier in Chapter 6, these values were proposed based on the results of other studies that can be applicable in Australia. Some of these values were proposed specifically based on the study area in Sydney and could be adjusted in accordance with the project objectives and desired targets set by planners and decision makers. Since all indicators in the model stem from an international survey with input from experts in all continents, future studies should be able to further improve this research by establishing national and/or international baselines.

Finally, this research tested the model on only one case study in Parramatta CBD, an area with dense urban structure and limited opportunity for increasing greenery. Therefore, future studies could apply this indicator-based model to different locations with various types of development (e.g. residential, from low to high density) and land-use patterns, which will enable inclusive comparisons in terms of achieving sustainability targets for effective and efficient GI types in relation to urban forms, patterns and surface characteristics as well as climatic zones. It is worth mentioning again that this proposed indicator-based model requires further improvement, adjustment and development in order to include other indicators that were excluded from the model, as well as to set up international baseline values.

231

REFERENCES

Abraham, A., Sommerhalder, K., & Abel, T. (2010). Landscape and well-being: A scoping study on the health-promoting impact of outdoor environments. International Journal of Public Health, 55(1), 59-69. Adams, L. W., & Dove, L. E. (1989). Wildlife reserves and corridors in the urban environment: a guide to ecological landscape planning and resource conservation. Columbia, Maryland: National Institute for Urban Wildlife. Agampatian, R. (2014). Using GIS to measure walkability: A Case study in New York City. School of Architecture and the Built Environment, Stockholm, Sweden. Aguaron, E., & McPherson, E. G. (2012). Comparison of methods for estimating carbon dioxide storage by Sacramento’s urban forest Carbon sequestration in urban ecosystems (pp. 43-71). Netherlands: Springer. Ahern, J. (1995). Greenways as a planning strategy. Landscape and urban planning, 33(1), 131-155. Ahern, J. (2007). Green Infrastructure for Cities: The Spatial Dimension. In V. Novotny & P. Brown (Eds.), Cities of the Future. London, UK: IWA Publishing. Ahern, J. (2010). Planning and design for sustainable and resilient cities: Theories, strategies, and best practices for green infrastructure. In V. Novotny, J. Ahern & P. Brown (Eds.), Water Centric Sustainable Communities: Planning, Retrofitting, and Building the Next Urban Environment (pp. 135-176). Hoboken, NJ, USA: John Wiley & Sons, Inc. Ahmed, K. (1994). A comparative analysis of the outdoor thermal environment of the urban vernacular and the contemporary development: case studies in Dhaka. Paper presented at the Proceedings of the 11 th PLEA International Conference, Dead Sea, 3-8 July 1994. Akbari, H. (2002). Shade trees reduce building energy use and CO2 emissions from power plants. Environmental pollution, 116, S119-S126. Akbari, H., Davis, S., Dorsano, S., Huang, J., & Winnett, S. (1992). Cooling our communities: a guidebook on tree planting and light-colored surfacing. Washington, D.C.: U.S. Environmental Protection Agency, Office of Policy Analysis, Climate Change Division. Akbari, H., Pomerantz, M., & Taha, H. (2001). Cool surfaces and shade trees to reduce energy use and improve air quality in urban areas. Solar energy, 70(3), 295-310. Akbari, H., Rosenfeld, A., Taha, H., & Gartland, L. (1996). Mitigation of summer urban heat islands to save electricity and smog. Paper presented at the 76th Annual Meteorological Society Meeting, Atlanta, GA. Akbari, H., & Taha, H. (1992). The impact of trees and white surfaces on residential heating and cooling energy use in four Canadian cities. Energy, 17(2), 141-149. Alberti, M., Marzluff, J. M., Shulenberger, E., Bradley, G., Ryan, C., et al. (2003). Integrating humans into ecology: opportunities and challenges for studying urban ecosystems. BioScience, 53(12), 1169-1179. Algert, S. J., Agrawal, A., & Lewis, D. S. (2006). Disparities in access to fresh produce in low-income neighborhoods in Los Angeles. American journal of preventive medicine, 30(5), 365-370. Ali-Toudert, F., Djenane, M., Bensalem, R., & Mayer, H. (2005). Outdoor thermal comfort in the old desert city of Beni-Isguen, Algeria. Climate Research, 28(3), 243-256.

232

Ali-toudert, F., & Mayer, H. (2006). Effects of street design on outdoor thermal comfort. Retrieved from http://www2.sci.u-szeged.hu/eghajlattan/baba/Ali-Toudert.pdf Ali-Toudert, F., & Mayer, H. (2007). Effects of asymmetry, galleries, overhanging facades and vegetation on thermal comfort in urban street canyons. Solar Energy, 81(6), 742- 754. Allen, R., Pereira, L., Raes, D., & Smith, M. (1998). Crop evapotranspiration—guidelines for computing crop water requirements: FAO Irrigation and Drainage paper no. 56. Food and Agricultural Organization of the United Nations, Rome, Italy, p 333. Allen, R. G., Tasumi, M., & Trezza, R. (2007). Satellite-based energy balance for mapping evapotranspiration with internalized calibration (METRIC)—Model. Journal of Irrigation and Drainage Engineering, 133(4), 380-394. Amaratunga, D., Baldry, D., Sarshar, M., & Newton, R. (2002). Quantitative and qualitative research in the built environment: application of mixed research approach. Work Study, 51(1), 17-31. Andersson, E., Barthel, S., Borgström, S., Colding, J., Elmqvist, T., et al. (2014). Reconnecting cities to the biosphere: Stewardship of green infrastructure and urban ecosystem services. Ambio, 43(4), 445-453. Armstrong, D. (2000). A survey of community gardens in upstate New York: Implications for health promotion and community development. Health & place, 6(4), 319-327. Arnfield, A. J. (2003). Two decades of urban climate research: A review of turbulence, exchanges of energy and water, and the urban heat island. International Journal of Climatology, 23(1), 1-26. ASHRAE, A. (2004). Standard 55-2004, Thermal environmental conditions for human occupancy. American Society of Heating, Refrigerating and Air-Conditioning Engineering, Atlanta, GA. ATHENA Institute. (2007). Applying Life Cycle Costing (LCC) to Roofing Investments: A Guide to Using Green Roofs for Healthy Cities GreenSave Calculator. North America. Auditor-General of Victoria. (2013). Managing Traffic Congestion. Australia. Austin, G. (2014). Green Infrastructure for Landscape Planning: Integrating Human and Natural Systems. New York: Routledge Press. Austin, S. B., Melly, S. J., Sanchez, B. N., Patel, A., Buka, S., et al. (2005). Clustering of fast-food restaurants around schools: a novel application of spatial statistics to the study of food environments. American Journal of Public Health, 95(9), 1575-1581. Australian Bureau of Statistics. (2011). 2074.0 - Census of Population and Housing: Mesh Block Counts. Retrieved July 2016: http://www.abs.gov.au/AUSSTATS/[email protected]/DetailsPage/2074.02011?OpenDocume nt Australian Bureau of Statistics. (2017). Monthly Rainfall (millimetres). Retrieved January 2017: http://www.bom.gov.au/climate/cdo/about/about-rain-data.shtml. Australian Government. (2013). Walking, Riding and Access to Public Transport ; Supporting active travel in Australian communities ministerial statement. Australia: Department of Infrastructure and Transport. Australian Government. (2016). Smart Cities Plan. Australia: Department of Prime Minister and Cabinet. Australian Institute of Landscape Architect (AILA). (2012). Adapting to Climate Change: Green Infrastructure. Canberra. Babbie, E. R. (1990). Survey Research Methods Wadsworth Pub. Co.

233

Babey, S. H., Hastert, T. A., & Brown, E. R. (2007). Teens living in disadvantaged neighborhoods lack access to parks and get less physical activity. UCLA Center for Health Policy Research, The California Endowment. Bagstad, K. J., Semmens, D. J., & Winthrop, R. (2013). Comparing approaches to spatially explicit ecosystem service modeling: A case study from the San Pedro River, Arizona. Ecosystem Services, 5, 40-50. Barton, H., & Grant, M. (2006). A health map for the local human habitat. The Journal for the Royal Society for the Promotion of Health, 126(6), 252-253. Barton, H., Grant, M., & Guise, R. (2006). Shaping neighbourhoods: For local health and global sustainability. London, UK: Taylor & Francis. Basinger, A. R., McKenney, C. B., & Auld, D. (2009). Competencies for a United States horticulture undergraduate major: A national Delphi study. HortTechnology, 19(2), 452-458. Bastiaanssen, W., Menenti, M., Feddes, R., & Holtslag, A. (1998). A remote sensing surface energy balance algorithm for land (SEBAL). 1. Formulation. Journal of hydrology, 212, 198-212. Battiata, J., Collins, K., Hirschman, D., & Hoffmann, G. (2010). The runoff reduction method. Journal of Contemporary Water Research & Education, 146(1), 11-21. Beatley, T. (2000). Green urbanism: Learning from European cities. Washington D.C.: Island Press. Beatley, T. (2012). Green cities of Europe: global lessons on green urbanism. Washington D.C.: Island Press. Beatley, T., & Newman, P. (2012). Green urbanism down under: Learning from sustainable communities in Australia. Washington D.C.: Island Press. Beattie, A., Hek, G., Ross, K., & Galvin, K. (2004). Future career pathways in nursing and midwifery. A Delphi survey of nurses and midwives in South West England. Nursing Times Research, 9(5), 348-364. Beckett, K. P., Freer Smith, P., & Taylor, G. (2000). Effective tree species for local air quality management. Journal of arboriculture, 26(1), 12-19. Beets, P. N., Kimberley, M. O., Oliver, G. R., Pearce, S. H., Graham, J. D., et al. (2012). Allometric equations for estimating carbon stocks in natural forest in New Zealand. Forests, 3(3), 818-839. Behar, S., Dates, G., & Byrne, J. (1997). Testing the waters: Chemical and physical vital signs of a river. Montpelier, VT: River Watch Network. Beier, P., Majka, D., & Jenness, J. (2007). Conceptual steps for designing wildlife corridors. Flagstaff: Northern Arizona University. Bell, S., & Morse, S. (2008). Sustainability indicators: measuring the immeasurable? UK: Earthscan. Benedict, M. A., & McMahon, E. T. (2002). Green Infrastructure: Smart Conservation for the 21st Century. Retrieved from http://www.sprawlwatch.org/greeninfrastructure.pdf Benedict, M. A., & McMahon, E. T. (2006). Green Infrastructure: Linking landscapes and communities. Washington, D.C.: Island Press. Benedict, M. A., & McMahon, E. T. (2012). Green infrastructure: linking landscapes and communities: Island Press. Black, D., & Black, J. (2009). A review of the urban development and transport impacts on public health with particular reference to Australia: trans-disciplinary research teams and some research gaps. International journal of environmental research and public health, 6(5), 1557-1596.

234

Blick, S. A., Kelly, F., & Skupien, J. J. (2004). New Jersey Stormwater Best Management Practices Manual. Retrieved from http://www.njstormwater.org/bmp_manual/NJ_SWBMP_4%20print.pdf. Boulanger, P.-M. (2008). Sustainable development indicators: a scientific challenge, a democratic issue. SAPI EN. S. Surveys and Perspectives Integrating Environment and Society(1.1). Bowler, D. E., Buyung-Ali, L., Knight, T. M., & Pullin, A. S. (2010). Urban greening to cool towns and cities: A systematic review of the empirical evidence. Landscape and urban planning, 97(3), 147-155. Boyd, B. L. (2003). Identifying competencies for volunteer administrators for the coming decade: A national Delphi study. Journal of Agricultural Education, 44(4), 47-56. Brack, C. L. (2002). Pollution mitigation and carbon sequestration by an urban forest. Environmental pollution, 116, S195-S200. Brady, A. B., Brake, D. R., & Starks, C. W. (2001). The Green infrastructure guide: Planning for a healthy urban and community forest. Trenton, NJ: Regional Planning Partnership. Breuste, J., Niemelä, J., & Snep, R. P. (2008). Applying landscape ecological principles in urban environments. Landscape Ecology, 23(10), 1139-1142. Brill, J. M., Bishop, M., & Walker, A. E. (2006). The competencies and characteristics required of an effective project manager: A web-based Delphi study. Educational technology research and development, 54(2), 115-140. Brown, B. B., Yamada, I., Smith, K. R., Zick, C. D., Kowaleski-Jones, L., et al. (2009). Mixed land use and walkability: Variations in land use measures and relationships with BMI, overweight, and obesity. Health & Place, 15(4), 1130-1141. Brown, G., Schebella, M. F., & Weber, D. (2014). Using participatory GIS to measure physical activity and urban park benefits. Landscape and Urban Planning, 121, 34- 44. Brown, G. G., & Reed, P. (2012). Social landscape metrics: Measures for understanding place values from public participation geographic information systems (PPGIS). Landscape Research, 37(1), 73-90. Brown, N., Crawford, I., Carley, S., & Mackway-Jones, K. (2006). A Delphi-based consensus study into planning for biological incidents. Journal of Public Health, 28(3), 238-241. Brownson, R. C., Baker, E. A., Housemann, R. A., Brennan, L. K., & Bacak, S. J. (2001). Environmental and policy determinants of physical activity in the United States. American journal of public health, 91(12), 1995-2003. Brunekreef, B., & Holgate, S. T. (2002). Air pollution and health. The Lancet, 360(9341), 1233-1242. Bryman, A. (2015). Social research methods (5 ed.). Oxford, UK: Oxford University Press. Budruk, M., & Phillips, R. (Eds.). (2011). Quality-of-life community indicators for parks, recreation and tourism management (Vol. 43): Springer Science+Business Media B.V. Burt, C. M., Mutziger, A. J., Allen, R. G., & Howell, T. A. (2005). Evaporation research: Review and interpretation. Journal of Irrigation and Drainage Engineering, 131(1), 37-58. CABE (Producer). (2011, March 2013) CABE Sustainable Places-Green Infrastructure. Commision for Architecture and Built Environment. retrieved from http://webarchive. nationalarchives.gov.uk/20110118095356/ http:/ www.cabe.org.uk/sustainable- places/green-infrastructure

235

Calabrese, J. M., & Fagan, W. F. (2004). A comparison‐shopper's guide to connectivity metrics. Frontiers in Ecology and the Environment, 2(10), 529-536. Cappiella, K., Claggett, S., Cline, K., Day, S., Galvin, M., et al. (2016a). Recommendations of the Expert Panel to Define BMP Effectiveness for Urban Tree Canopy Expansion. Retrieved December, 2016, from http://www.chesapeakebay.net/channel_files/23663/urban_tree_canopy_ep_report_for _wtwg_review_081816_clean.pdf Cappiella, K., Claggett, S., Cline, K., Day, S., Galvin, M., et al. (2016b). Recommendations of the Expert Panel to Define BMP Effectiveness for Urban Tree Canopy Expansion. Carter, J. G. (2011). Climate change adaptation in European cities. Current opinion in environmental sustainability, 3(3), 193-198. Carter, T., & Butler, C. (2008). Ecological impacts of replacing traditional roofs with green roofs in two urban areas. Cities and the Environment (CATE), 1(2), 9. Castro, F. G., Kellison, J. G., Boyd, S. J., & Kopak, A. (2010). A methodology for conducting integrative mixed methods research and data analyses. Journal of Mixed Methods Research, 4(4), 342-360. Center for Neighbourhood Technology (CNT). (2010). The value of green infrastructure: A guide to recognizing its economic, environmental and social benefits. Chicago: Center for Neighborhood Technology. Center for Urban Forest Research (CUFR). (2008). CUFR Tree Carbon Calculator (CTCC) https://www.fs.usda.gov/ccrc/tools/tree-carbon-calculator-ctcc Chen, J. M., Plummer, P. S., Rich, M., Gower, S. T., & Norman, J. M. (1997). Leaf area index measurements. Journal of Geophysical Research, 102(D24), 29-429. Cheng, C., Cheung, K. K., & Chu, L. (2010). Thermal performance of a vegetated cladding system on facade walls. Journal of Building and Environment, 45(8), 1779-1787. Cheng, Y., Niu, J., & Gao, N. (2012). Thermal comfort models: A review and numerical investigation. Building and Environment, 47, 13-22. Christopherson, R. W. (2007). Elemental Geosystems (5 ed.). NJ, USA: Prentice Hall. City of Melbourne. (2011). Urban Forest Strategy: Making a great city greener 2012-2032. Melbourne, Australia: City of Melbourne. City of Sydney. (2012). Decentralised water master plan; WSUD and stormwater infrastructure report. Retrieved from http://www.cityofsydney.nsw.gov.au/__data/assets/pdf_file/0015/130515/130218_EC _ITEM02_ATTACHMENTD6.PDF. City of Sydney. (2013). Urban Forest Strategy. Sydney, Australia: City of Sydney. Cohen, D. A., McKenzie, T. L., Sehgal, A., Williamson, S., Golinelli, D., et al. (2007). Contribution of public parks to physical activity. American Journal of Public Health, 97(3), 509-514. Connellan, G. (2002). Efficient irrigation: A reference manual for turf and landscape. Melbourne, Australia: Brunley College, University of Melbourne. Conway, E. D., & The Maryland Space Grant Consortium. (1997). An introduction to satellite image interpretation. Baltimore and London: The John Hopkins University Press. Coombes, E., Jones, A. P., & Hillsdon, M. (2010). The relationship of physical activity and overweight to objectively measured green space accessibility and use. Social Science & Medicine, 70(6), 816-822. Costanza, R., d’Arge, R., de Groot, R., Farber, S., Grasso, M., et al. (2017). The value of the world’s ecosystem services and natural capital (1997). In P. Newell & J. T. Roberts

236

(Eds.), The Globalization and Environment Reader (pp. 117-133). UK: John Wiley & Sons, Ltd. Costello, L. R., & Jones, K. S. (1994). WUCOLS, Water use classification of landscape species: A guide to the water needs of landscape plants. California, USA: California Department of Water Resources. Costello, L. R., & Jones, K. S. (2014). WUCOLS IV, Water use classification of landscape species. California, USA: University of California, Cooperative Extension. Costello, L. R., Matheny, N. P., Clark, J. R., & Jones, K. S. (2000). A guide to estimating irrigation water needs of landscape plantings in California. California, USA: University of California Cooperative Extension, California Department of Water Resources. Coutts, A. M., Beringer, J., & Tapper, N. J. (2007). Impact of increasing urban density on local climate: Spatial and temporal variations in the surface energy balance in Melbourne, Australia. Journal of Applied Meteorology and Climatology, 46(4), 477- 493. Coutts, A. M., Tapper, N. J., Beringer, J., Loughnan, M., & Demuzere, M. (2012). Watering our cities: the capacity for water sensitive urban design to support urban cooling and improve human thermal comfort in the Australian context. Progress in Physical Geography, 0309133312461032. Coutts, A. M., White, E. C., Tapper, N. J., Beringer, J., & Livesley, S. J. (2016). Temperature and human thermal comfort effects of street trees across three contrasting street canyon environments. Theoretical and Applied Climatology, 124(1-2), 55-68. Coutts, C. (2016). Green infrastructure and public health. London, UK: Routledge. Coutts, C., Forkink, A., & Weiner, J. (2014). The portrayal of natural environment in the evolution of the ecological public health paradigm. International Journal of Environmental Research and Public Health, 11(1), 1005-1019. Creswell, J. W. (2003). Research design: Qualitative, quantitative, and mixed methods approaches (2 ed.). London, UK: Sage Publications. Creswell, J. W., & Garrett, A. L. (2008). The" movement" of mixed methods research and the role of educators. South African Journal of Education, 28(3), 321-333. Currie, B. A., & Bass, B. (2008). Estimates of air pollution mitigation with green plants and green roofs using the UFORE model. Urban Ecosystems, 11(4), 409-422. Dale, V. H. (1997). The relationship between land-use change and climate change. Ecological Applications 7, 753–769. Davies, C., Macfarlane, R., & Roe, M. (2006). Green infrastructure planning guide project. Annfield Plain, UK: University of Northumbria, North East Community Forests, University of Newcastle, Countryside Agency, English Nature, Forestry Commission, Groundwork Trusts. Davies, M. (2011). Concept mapping, mind mapping and argument mapping: what are the differences and do they matter? Higher education, 62(3), 279-301. de Groot, R., Brander, L., Van Der Ploeg, S., Costanza, R., Bernard, F., et al. (2012). Global estimates of the value of ecosystems and their services in monetary units. Ecosystem Services, 1(1), 50-61. de Groot, R. S., Alkemade, R., Braat, L., Hein, L., & Willemen, L. (2010). Challenges in integrating the concept of ecosystem services and values in landscape planning, management and decision making. Ecological Complexity, 7(3), 260-272. de Groot, R. S., Wilson, M. A., & Boumans, R. M. (2002). A typology for the classification, description and valuation of ecosystem functions, goods and services. Ecological Economics, 41(3), 393-408.

237

Derkzen, M. L., van Teeffelen, A. J., & Verburg, P. H. (2017). Green infrastructure for urban climate adaptation: How do residents’ views on climate impacts and green infrastructure shape adaptation preferences? Landscape and Urban Planning, 157, 106-130. Desha, C., Reeve, A. C., Newman, P. W., & Beatley, T. (2016). Urban nature for resilient and liveable cities. Smart and Sustainable Built Environment, 5(1). Di Leo, N., Escobedo, F. J., & Dubbeling, M. (2016). The role of urban green infrastructure in mitigating land surface temperature in Bobo-Dioulasso, Burkina Faso. Environment, Development and Sustainability, 18(2), 373-392. Díaz, S., Quétier, F., Cáceres, D. M., Trainor, S. F., Pérez-Harguindeguy, N., et al. (2011). Linking functional diversity and social actor strategies in a framework for interdisciplinary analysis of nature's benefits to society. Proceedings of the National Academy of Sciences, 108(3), 895-902. Diener Ed, Suh, E. M., Lucas, R. E., & Smith, H. L. (1999). Subjective well-being: Three decades of progress. Psychological Bulletin, 125(2), 276-302. Dige, G. (2011). Green infrastructure and territorial cohesion. The concept of green infrastructure and its integration into policies using monitoring systems: Technical Report 18. European Environment Agency, Copenhagen, Denmark.[online] URL: http://www. eea. europa. eu/publications/green-infrastructure-and-territorial-cohesion. Dimoudi, A., & Nikolopoulou, M. (2003). Vegetation in the urban environment: microclimatic analysis and benefits. Energy and buildings, 35(1), 69-76. Ding, D., Sallis, J. F., Kerr, J., Lee, S., & Rosenberg, D. E. (2011). Neighborhood environment and physical activity among youth: A review. American Journal of Preventive Medicine, 41(4), 442-455. Dizdaroglu, D. (2013). A micro-level indexing model for the assessment of sustainable urban ecosystems. Queensland University of Technology, Brisbane, Queensland. Dobbins, T. R. (1999). Clinical experiences for agricultural teacher education programs in North Carolina, South Carolina, and Virginia. Virginia Polytechnic Institute and State University, Virginia, USA. Donovan, G. H., & Butry, D. T. (2009). The value of shade: Estimating the effect of urban trees on summertime electricity use. Energy and Buildings, 41(6), 662-668. Douglas, I. (2012). Urban ecology and urban ecosystems: Understanding the links to human health and well-being. Current Opinion in Environmental Sustainability, 4(4), 385- 392. Dutt, A. K., Noble, A. G., Costa, F. J., Thakur, R. R., Thakur, S. K., et al. (Eds.). (2015). Spatial diversity and dynamics in resources and urban development (Vol. 1): Springer Netherlands. Dwyer, J. F., McPherson, E. G., Schroeder, H. W., & Rowntree, R. A. (1992). Assessing the benefits and costs of the urban forest. Journal of Arboriculture, 18, 227-227. Ecologically Sustainable Development Steering Committee. (1992). National Strategy for Ecologically Sustainable Development - Part 1 Introduction. Australia: Australian Government, Department of the Environment and Energy. Edwards, C., & Miller, M. (2001). PLOAD Version 3.0 User’s Manual; An ArcView GIS Tool to Calculate Nonpoint Sources of Pollution in Watershed and Stormwater Projects. Washington, D.C.: US EPA. Eisenman, T. S. (2013). Frederick Law Olmsted, green infrastructure, and the evolving city. Journal of Planning History, 12(4), 287-311. Ekkel, E. D., & de Vries, S. (2017). Nearby green space and human health: Evaluating accessibility metrics. Landscape and Urban Planning, 157, 214-220.

238

Ellaway, A., Macintyre, S., & Bonnefoy, X. (2005). Graffiti, greenery, and obesity in adults: Secondary analysis of European cross sectional survey. BMJ, 331(7517), 611-612. Ely, M., & Pitman, S. (2012). Green infrastructure: Life support for human habitats. Ely, M., & Pitman, S. (2014). Green infrastructure: Life support for human habitats. Enlow, C. (2002). Rethinking streets as parks-Seattle's open-spase strategy shifts its emphasis away from traditional parks to an evolving urban ecosystem that focuses on city streets. Landscape Architecture, 92(9), 52-58. ENSPEC (Producer). (2012, September) i-Tree Australia. Arboriculture Australia. retrieved from http://arboriculture.org.au/i-Tree-Australia Environment and Planning References Committee. (2012). Inquiry Into Environmental Design and Public Health in Victoria (No. 1). Melbourne: Government Printer for the State of Victoria. Environmental Protection Agency(EPA) (Producer). (2012, March) Green Infrastructure Modeling Tools. United States Environmental Protection Agency. retrieved from http://water.epa.gov/infrastructure/greeninfrastructure/gi_modelingtools.cfm Escobedo, F., Varela, S., Zhao, M., Wagner, J. E., & Zipperer, W. (2010). Analyzing the efficacy of subtropical urban forests in offsetting carbon emissions from cities. Environmental Science & Policy, 13(5), 362-372. European Commision. (2012). The Multifunctionality of Green Infrastructure: Science for Environment Policy. European Commission. (2009). Towards a green infrastructure for Europe: Developing new concepts for integrating of Natura 2000 network into the broader countryside EC study. ENV.B.2/SER/2007/0076. European Environment Agency. (2011). Green Infrastructure and Territorial Cohesion: The Concept of Green Infrastructure and Its Integration Into Policies Using Monitoring Systems: Publications Office of the European Union. Evenson, K. R., Wen, F., Hillier, A., & Cohen, D. A. (2013). Assessing the contribution of parks to physical activity using GPS and accelerometry. Medicine and Science in Sports and Exercise, 45(10), 1981-1987. eWater (Producer). (2013) Evolving water management. retrieved from http://www.ewater.com.au/products/ewater-toolkit/urban-tools/music/ Fam, D., Mosley, E., Lopes, A., Mathieson, L., Morison, J., et al. (2008). Irrigation of Urban Green Spaces: A review of the environmental, social and economic benefits. CRC for Irrigation Futures Technical Report(04/08). Farr, D. (2011). Sustainable urbanism: Urban design with nature. Hoboken, New Jersey: John Wiley & Sons. Favilli, F., Hoffmann, C., Elmi, M., Ravazzoli, E., & Streifeneder, T. (2015). The BioREGIO Carpathians project: Aims, methodology and results from the continuity and connectivity analysis. Nature Conservation, 11, 95. Fisher, B., Turner, R. K., & Morling, P. (2009). Defining and classifying ecosystem services for decision making. Ecological economics, 68(3), 643-653. Fletcher, T., Duncan, H., Poelsma, P., & Lloyd, S. (2004). Stormwater Flow and Quality, and the Effectiveness of Non-Proprietary Stormwater Treatment Measures: A Review and Gap Analysis. . Australia. Flynn, K. M. (2011). Evaluation of green infrastructure practices using life cycle assessment. Villanova, Pennsylvania: Villanova University: College of Engineering. Forman, R. T. (1995). Landscape Mosaics: The ecology of landscape and regions. NY: University of Cambridge.

239

Forsyth, A., Oakes, J. M., Schmitz, K. H., & Hearst, M. (2007). Does residential density increase walking and other physical activity? Urban Studies, 44(4), 679-697. Foster, J., Lowe, A., & Winkelman, S. (2011). The value of green infrastructure for urban climate adaptation. Washington, D.C.: Center for Clean Air Policy. Francis, J., Wood, L. J., Knuiman, M., & Giles-Corti, B. (2012). Quality or quantity? Exploring the relationship between Public Open Space attributes and mental health in Perth, Western Australia. Social Science & Medicine, 74(10), 1570-1577. Frank, L. D., Saelens, B. E., Chapman, J., Sallis, J. F., Kerr, J., et al. (2012). Objective assessment of obesogenic environments in youth: geographic information system methods and spatial findings from the neighborhood impact on kids study. American Journal of Preventive Medicine, 42(5), e47-e55. Frank, L. D., Sallis, J. F., Conway, T. L., Chapman, J. E., Saelens, B. E., et al. (2006). Many pathways from land use to health: associations between neighborhood walkability and active transportation, body mass index, and air quality. Journal of the American Planning Association, 72(1), 75-87. Freeman, H. L. (1984). Mental health and the environment: Churchill Livingstone. Fuller, R. A., & Gaston, K. J. (2009). The scaling of green space coverage in European cities. Biology letters, 5(3), 352-355. Garrigues, S., Shabanov, N., Swanson, K., Morisette, J., Baret, F., et al. (2008). Intercomparison and sensitivity analysis of Leaf Area Index retrievals from LAI- 2000, AccuPAR, and digital hemispherical photography over croplands. Agricultural and Forest Meteorology, 148(8), 1193-1209. Gasparatos, A. (2010). Embedded value systems in sustainability assessment tools and their implications. Journal of Environmental Management, 91(8), 1613-1622. Gay, L. R., & Diehl, P. L. (1992). Research methods for business and management. New York and Toronto: MacMillan Coll Div Publishing George, D., & Mallery, P. (2011). Using SPSS for Windows step by step: A simple guide and reference 18.0 Update (11 ed.). New Jersey, USA: Allyn & Bacon/Pearson. GHD. (2011). Report for urban heat island effect, mitigation strategies and planning policy approaches. Melbourne, Vic: City of Melbourne. Ghosh, S. (2014). Measuring sustainability performance of local food production in home gardens. Local Environment, 19(1), 33-55. Gill, S. E., Handley, J. F., Ennos, A. R., & Pauleit, S. (2007). Adapting cities for climate change: The role of the green infrastructure. Built Environment, 33(1), 115-133. Godden, B. (2004). Sample size formulas. Retrieved April 2016: http://www.williamgodden.com/samplesizeformula.pdf Goodman, L. A. (2011). Comment: On respondent‐driven sampling and snowball sampling in hard‐to‐reach populations and snowball sampling not in hard‐to‐reach populations. Sociological Methodology, 41(1), 347-353. Goodman, M. B., & Santos, G. J. (2006). Card sort technique as a qualitative substitute for quantitative exploratory factor analysis. Corporate Communications: An International Journal, 11(3), 288-302. Government Architect’s Office NSW Public Works. (2016). Paramatta Ways, Implementing Sydney's Green Grid. Sydney, Australia: Parramatta City Council. Gower, S. T., & Norman, J. M. (1991). Rapid estimation of leaf area index in conifer and broad‐leaf plantations. Ecology, 72(5), 1896-1900. Grace, P., & Basso, B. (2012). Offsetting greenhouse gas emissions through biological carbon sequestration in North Eastern Australia. Agricultural Systems, 1(105), 1-6.

240

Graham, S., & Marvin, S. (2001). Splintering urbanism: Networked infrastructures, technological mobilities and the urban condition. London, UK: Routledge. Grahn, P., & Stigsdotter, U. A. (2003). Landscape planning and stress. Urban forestry & urban greening, 2(1), 1-18. Gray, D. E. (2009). Doing research in the real world (2 ed.). London, UK: SAGE Publication Ltd. Greater London Authority. (2012). Green infrastructure and open environments: The all London green grid supplementary planning guidance. Retrieved from https://www.london.gov.uk/sites/default/files/algg_spg_mar2012.pdf. Green Building Council of Australia. (2014). Green Star: Potable water calculator guide Retrieved July 2016: https://www.gbca.org.au/uploads/233/35817/Potable%20Water%20Calculator%20Gu ide%20v1%2020102014.pdf Greene, J. C. (2008). Is mixed methods social inquiry a distinctive methodology? Journal of Mixed Methods Research, 2(1), 7-22. Grimm, N. B., Morgan Grove, J., Pickett, S. T., & Redman, C. L. (2000). Integrated Approaches to Long-TermStudies of Urban Ecological Systems: Urban ecological systems present multiple challenges to ecologists—pervasive human impact and extreme heterogeneity of cities, and the need to integrate social and ecological approaches, concepts, and theory. BioScience, 50(7), 571-584. Hahn, J. T. (1984). Tree volume and biomass equations for the Lake States. United States U.S. Department of Agriculture, Forest Service, North Central Forest Experiment Station. Hansen, M., DeFries, R., Townshend, J., Carroll, M., Dimiceli, C., et al. (2003). Global percent tree cover at a spatial resolution of 500 meters: First results of the MODIS vegetation continuous fields algorithm. Earth Interactions, 7(10), 1-15. Hansen, R., & Pauleit, S. (2014). From multifunctionality to multiple ecosystem services? A conceptual framework for multifunctionality in green infrastructure planning for urban areas. Ambio, 43(4), 516-529. Hargreaves, G. H., & Samani, Z. A. (1982). Estimating potential evapotranspiration. Journal of the Irrigation and Drainage Division, 108(3), 225-230. Hargreaves, G. H., & Samani, Z. A. (1985). Reference crop evapotranspiration from temperature. Applied engineering in agriculture, 1(2), 96-99. Harris, R. W., Clark, J. R., & Matheny, N. P. (2004). Arboriculture: Integrated management of landscape trees, shrubs, and vines (4th ed.). Engelwood Cliffs, New Jersey, USA: Prentice-Hall International. Harrison, J. (2005). Future personal attributes and job competencies needed by the Texas Department of State Health Services (DSHS), state hospital section, registered nurse managers: A Delphi study. Texas A&M University, Texas, USA. Hartig, T., Mitchell, R., De Vries, S., & Frumkin, H. (2014). Nature and health. Annual Review of Public Health, 35, 207-228. Haug, A., Brand-Miller, J. C., Christophersen, O. A., McArthur, J., Fayet, F., et al. (2007). A food lifeboat: Food and nutrition considerations in the event of a pandemic or other catastrophe. Medical Journal of Australia, 187(11), 674-675. Heckathorn, D. D. (1997). Respondent-driven sampling: A new approach to the study of hidden populations. Social Problems, 44(2), 174-199. Heckathorn, D. D. (2002). Respondent-driven sampling II: Deriving valid population estimates from chain-referral samples of hidden populations. Social Problems, 49(1), 11-34.

241

Heinberg, R. (2010). Turning the corner on growth. Solutions, 1(1), 14-16. Heink, U., & Kowarik, I. (2010). What are indicators? On the definition of indicators in ecology and environmental planning. Ecological Indicators, 10(3), 584-593. Heisler, G. M. (1986). Energy savings with trees. Journal of Aboriculture, 12(5), 113-125. Henwood, K. (2002). Environment and health: Is there a role for environmental and countryside agencies in promoting benefits to health? Health Development Agency, 1- 16. Hirabayashi, S. (2013a). i-Tree Eco precipitation interception model descriptions. Version 1.3. Washington, D.C.: U.S. Department of Agriculture, Forest Service and Kent: Davey Tree Expert Company. Hirabayashi, S. (2013b). i-Tree Streets/Design/Eco rainfall interception model comparisons. Washington, D.C.: U.S. Department of Agriculture, Forest Service and Kent: Davey Tree Expert Company. Hirschman, D., Collins, K., & Schueler, T. (2008). Technical memorandum: The runoff reduction method. Holling, C. S. (1973). Resilience and stability of ecological systems. Annual review of ecology and systematics, 4(1), 1-23. Hostetler, M., Allen, W., & Meurk, C. (2011). Conserving urban biodiversity? Creating green infrastructure is only the first step. Landscape and Urban Planning, 100(4), 369-371. Hough, M. (2004). Cities and natural processes: A basis for sustainability. London: Routledge. Howes, R., & Robinson, H. (2005). Infrastructure for the built environment: Global procurement strategies: Routledge. Huang, L., Li, J., Zhao, D., & Zhu, J. (2008). A fieldwork study on the diurnal changes of urban microclimate in four types of ground cover and urban heat island of Nanjing, China. Building and environment, 43(1), 7-17. Hynicka, J., & Divers, M. (2016). Relative reductions in non-point source pollution loads by urban trees. Retrieved from http://www.chesapeakebay.net/channel_files/23663/urban_tree_canopy_ep_report_for _wtwg_review_081816_clean.pdf Ichinose, T., Shimodozono, K., & Hanaki, K. (1999). Impact of anthropogenic heat on urban climate in Tokyo. Atmospheric Environment, 33(24), 3897-3909. Jacobson, B., & Peres-Neto, P. R. (2010). Quantifying and disentangling dispersal in metacommunities: how close have we come? How far is there to go? Landscape Ecology, 25(4), 495-507. Jamieson, S. (2004). Likert scales: how to (ab) use them. Medical education, 38(12), 1217- 1218. Jarvis, P., & Leverenz, J. (1983). Productivity of temperate, deciduous and evergreen forests Physiological plant ecology IV (pp. 233-280): Springer. Jax, K. (2003). Die Funktion biologischer Vielfalt. Naturschutzbegründungen. Bundesamt für Naturschutz/Landwirtschaftsverlag GmbH, Bonn und Münster, 149-174. Jenkins, J. C., Chojnacky, D. C., Heath, L. S., & Birdsey, R. A. (2003). National-scale biomass estimators for United States tree species. Forest Science, 49(1), 12-35. Johansson, E. (2006). Influence of urban geometry on outdoor thermal comfort in a hot dry climate: a study in Fez, Morocco. Building and environment, 41(10), 1326-1338. Johansson, E., & Emmanuel, R. (2006). The influence of urban design on outdoor thermal comfort in the hot, humid city of Colombo, Sri Lanka. International journal of biometeorology, 51(2), 119-133.

242

Johnson, R. B., Onwuegbuzie, A. J., & Turner, L. A. (2007). Toward a definition of mixed methods research. Journal of mixed methods research, 1(2), 112-133. Joint Research Centre-European Commission. (2008). Handbook on constructing composite indicators: Methodology and User guide: OECD publishing. Jonckheere, I., Fleck, S., Nackaerts, K., Muys, B., Coppin, P., et al. (2004). Review of methods for in situ leaf area index determination: Part I. Theories, sensors and hemispherical photography. Agricultural and forest meteorology, 121(1), 19-35. Kambites, C., & Owen, S. (2006). Renewed prospects for green infrastructure planning in the UK 1. Planning, Practice & Research, 21(4), 483-496. Kaplan, R. (1973). Some psychological benefits of gardening. Environment and behavior, 5(2), 145. Kaplan, R., & Kaplan, S. (1989). The experience of nature: A psychological perspective: CUP Archive. Kent, J., Thompson, S., & Jalaludin, B. (2011). Healthy Built Environments: A review of the literature. Australia: Healthy Built Evironments Program, City Futures Research Centre, The University of New South Wales. Khan, S., & Byrnes, K. (2016). How should recreational water quality in the Parramatta River be assessed? Australia: Sydney, Parramatta City Council,. Killicoat, P., Puzio, E., & Stringer, R. (2002). The economic value of trees in urban areas: estimating the benefits of Adelaide's street trees. TREENET, 90. Killus, J., Meyer, J., Durran, D., Anderson, G., & Jerskey, T. (1984). Continued research in mesoscale air-pollution simulation modeling. Volume 5. Refinements in numerical analysis, transport, chemistry, and pollutant removal. Final report, October 1979- July 1982.[AIRSHED model and SHASTA method]: Systems Applications, Inc., San Rafael, CA (USA). Kim, J., & Kaplan, R. (2004). Physical and psychological factors in sense of community new urbanist Kentlands and nearby Orchard Village. Environment and behavior, 36(3), 313-340. Klein, A.-M., Vaissiere, B. E., Cane, J. H., Steffan-Dewenter, I., Cunningham, S. A., et al. (2007). Importance of pollinators in changing landscapes for world crops. Proceedings of the Royal Society of London B: Biological Sciences, 274(1608), 303- 313. Kortenkamp, K. V., & Moore, C. F. (2001). Ecocentrism and anthropocentrism: Moral reasoning about ecological commons dilemmas. Journal of Environmental Psychology, 21(3), 261-272. Kothari, C. R. (2004). Research methodology: Methods and techniques: New Age International. Kucharik, C., Norman, J., Murdock, L. M., & Gower, S. (1997). Characterizing canopy nonrandomness with a multiband vegetation imager (MVI). Journal of Geophysical Research: Atmospheres, 102(D24), 29455-29473. Kuo, F. E., Bacaicoa, M., & Sullivan, W. C. (1998). Transforming inner-city landscapes trees, sense of safety, and preference. Environment and behavior, 30(1), 28-59. Kuo, F. E., & Sullivan, W. C. (2001). Aggression and violence in the inner city effects of environment via mental fatigue. Environment and behavior, 33(4), 543-571. Kushari, B., Mulyono, A. T., & Hendratno, A. (2015). Least-cost Path and Multi-criteria Decision Analyses on the Sumatera High-Grade Highway Route Plans. Proceedings of the Eastern Asia Society for Transportation Studies, 11.

243

Lafortezza, R., Davies, C., Sanesi, G., & Konijnendijk, C. C. (2013). Green Infrastructure as a tool to support spatial planning in European urban regions. iForest-Biogeosciences and Forestry, 6(3), 102-108. Lal, R., & Augustin, B. (2011). Carbon sequestration in urban ecosystems: Springer Science & Business Media. Lang, A. (1986). Leaf-area and average leaf angle from transmission of direct sunlight. Australian Journal of Botany, 34(3), 349-355. Lee, A. C., & Maheswaran, R. (2011). The health benefits of urban green spaces: a review of the evidence. Journal of public health, 33(2), 212-222. Lee, J. G., & Heaney, J. P. (2003). Estimation of urban imperviousness and its impacts on storm water systems. Journal of Water Resources Planning and Management, 129(5), 419-426. Leibold, M. A., Holyoak, M., Mouquet, N., Amarasekare, P., Chase, J. M., et al. (2004). The metacommunity concept: a framework for multi‐scale community ecology. Ecology letters, 7(7), 601-613. Lein, J. K. (2008). Integrated Environmental Planning: A Landscape Synthesis: John Wiley & Sons. Leopold, A. (1949). A sand county almanac, with essays on conservation from Roun River. London: Oxford University Press. Levasseur, M. (2014). Cycling aspects of Austroads guides (No. 1925037592). Australia: Austroads. Likert, R. (1967). The method of constructing and attitude scale. Methods and Trchniques in Business Research, 54. Lin, T.-P., Ho, Y.-F., & Huang, Y.-S. (2007). Seasonal effect of pavement on outdoor thermal environments in subtropical Taiwan. Building and environment, 42(12), 4124-4131. Lin, T.-P., Matzarakis, A., & Hwang, R.-L. (2010). Shading effect on long-term outdoor thermal comfort. Building and Environment, 45(1), 213-221. Lindberg, F., & Grimmond, C. (2011). The influence of vegetation and building morphology on shadow patterns and mean radiant temperatures in urban areas: model development and evaluation. Theoretical and Applied Climatology, 105(3-4), 311-323. Liu, C., Zhang, X., & Zhang, Y. (2002). Determination of daily evaporation and evapotranspiration of winter wheat and maize by large-scale weighing lysimeter and micro-lysimeter. Agricultural and Forest Meteorology, 111(2), 109-120. Livesley, S. (2010). Energy saving benefits of shade trees in relation to water use. Paper presented at the TREENET Proceedings of the 10th National Street Tree Symposium September 2010. Loughnan, M. E., Nicholls, N., & Tapper, N. J. (2010). When the heat is on: Threshold temperatures for AMI admissions to hospital in Melbourne Australia. Applied Geography, 30(1), 63-69. Lovell, S. T., & Taylor, J. R. (2013). Supplying urban ecosystem services through multifunctional green infrastructure in the United States. Landscape ecology, 28(8), 1447-1463. Maas, J., Van Dillen, S. M., Verheij, R. A., & Groenewegen, P. P. (2009). Social contacts as a possible mechanism behind the relation between green space and health. Health & place, 15(2), 586-595. Macfarlane, C., Grigg, A., & Evangelista, C. (2007). Estimating forest leaf area using cover and fullframe fisheye photography: thinking inside the circle. Agricultural and Forest Meteorology, 146(1), 1-12.

244

MacFarlane, R. (2007). Multi-functional Landscape: conceptual and planning issues for the countryside. In J. Benson & M. Roe (Eds.), Landscape & Sustainability (2nd ed.). New York: Routledge Press. Macintyre, S., Ellaway, A., & Cummins, S. (2002). Place effects on health: how can we conceptualise, operationalise and measure them? Social Science & Medicine, 55(1), 125-139. MacKinnon, B. (2007). Ethics: Theory and Contemporary Issues: Thomson/Wadsworth. Maco, S. E., & McPherson, E. G. (2003). A practical approach to assessing structure, function, and value of street tree populations in small communities. Margolis, L., & Robinson, A. (2007). Living systems: innovative materials and technologies for landscape architecture: Walter de Gruyter. Markart, G., Klebinder, K., & Kohl, B. (2008). Indicators for evaluation of surface runoff potential in critical land use areas. Paper presented at the Proceedings of the conference MONITOR/08–Systems behind a safer environment. International Conference on Monitoring Methods, 21st/22nd Feb. Markart, G., Kohl, B., Kirnbauer, R., Pirkl, H., Bertle, H., et al. (2006). Surface runoff in a torrent catchment area in Middle Europe and its prevention. Geotechnical & Geological Engineering, 24(5), 1403-1424. Marritz, L. (2011). Storm Water Quantity and Rate Control Benefits of Trees in Un compacted Soil. Green Infrastructure, 2016, from http://www.deeproot.com/blog/blog-entries/stormwater-quantity-and-rate-control- benefits-of-trees-in-uncompacted-soil Masadeh, M. A. (2012). Linking philosophy, methodology, and methods: Toward mixed model design in the hospitality industry. European Journal of Social Sciences, 28(1), 128-137. Mather, I., & Morton, A. (2008). Street Tree Design Guidelines. Matzarakis, A., De Rocco, M., & Najjar, G. (2009). Thermal bioclimate in Strasbourg-the 2003 heat wave. Theoretical and Applied Climatology, 98(3-4), 209-220. Mayer, A. L. (2008). Strengths and weaknesses of common sustainability indices for multidimensional systems. Environment international, 34(2), 277-291. Mazza, L., Bennett, G., De Nocker, L., Gantioler, S., Losarcos, L., et al. (2011). Green infrastructure implementation and efficiency. Institute for European Environmental Policy, Brussels. McDonald, L., Allen, W., Benedict, M. A., & O'Conner, K. (2005). Green infrastructure evaluation frameworks. Journal of Conservation Planning, 1(1), 12-43. McDonnell, M. J., & Pickett, S. T. (1990). Ecosystem structure and function along urban‐ rural gradients: an unexploited opportunity for ecology. Ecology, 71(4), 1232-1237. McGarigal, K. (2002). Landscape pattern metrics. Encyclopedia of environmetrics. McGarigal, K., Cushman, S. A., Neel, M. C., & Ene, E. (2002). FRAGSTATS: spatial pattern analysis program for categorical maps. McHale, M. R., Burke, I. C., Lefsky, M. A., Peper, P. J., & McPherson, E. G. (2009). Urban forest biomass estimates: is it important to use allometric relationships developed specifically for urban trees? Urban Ecosystems, 12(1), 95-113. McHarg, I. L. (1969). Design with Nature. Philadelphia, Pennsylvania: Natural History Press. McKinney, M. L. (2002). Urbanization, biodiversity, and conservation the impacts of urbanization on native species are poorly studied, but educating a highly urbanized human population about these impacts can greatly improve species conservation in all ecosystems. BioScience, 52(10), 883-890.

245

McPherson, E., Simpson, J., Peper, P., & Aguaron, E. (2008). Urban forestry and climate change. Albany, CA: USDA Forest Service, Pacific Southwest Research Station. Link. McPherson, E. G. (1994). Using urban forests for energy efficiency and carbon storage. Journal of Forestry;(United States), 92(10). McPherson, E. G. (1998). Atmospheric carbon dioxide reduction by Sacramento's urban forest. Journal of Arboriculture, 24, 215-223. McPherson, E. G. (2003). A Benefit–Cost Analysis Of Ten Street Tree Species In Modesto, California, U.S. Arboriculture. McPherson, E. G. (2010). Tools for valuing tree and park services. McPherson, E. G., Nowak, N., Heisler, G., Grimmond, S., Souch, C., et al. (1997). Quantifying urban forest structure, function and value: The Chicago Urban Forest Climate Project. Urban Ecosystems, 1, 49-61. McPherson, E. G., & Rowntree, R. A. (1993). Energy conservation potential of urban tree planting. Journal of Arboriculture, 19, 321-321. McPherson, E. G., & Simpson, J. (2009). Urban Forests and Climate Change. Retrieved March 2013: http://www.fed.us/ccrc/topics/urban-forests/ McPherson, E. G., & Simpson, J. R. (1999). Guidelines for calculating carbon dioxide reductions through urban forestry programs. Gen. Tech. Rep. PSW-171. Albany, CA: US Department of Agriculture, Forest Service, Pacific Southwest Research Station. McPherson, E. G., van Doorn, N. S., & Peper, P. J. (2016). Urban tree database and allometric equations. McPherson, E. G., Xiao, Q., & Aguaron, E. (2013). A new approach to quantify and map carbon stored, sequestered and emissions avoided by urban forests. Landscape and Urban Planning, 120, 70-84. McRae, B. H., Dickson, B. G., Keitt, T. H., & Shah, V. B. (2008). Using circuit theory to model connectivity in ecology, evolution, and conservation. Ecology, 89(10), 2712- 2724. Mell, I. (2008). Green infrastructure: concepts and planning. FORUM Ejournal, 8(1), 69-80. Mell, I. (2010). Green infrastructure: concepts, perceptions and its use in spatial planning. UNiversity of Newcastle. Mell, I. (2016). Global Green Infrastructure: Lessons for Successful Policy-making, Investment and Management: Routledge. Millennium Ecosystem Assessment. (2003). Ecosystems and human well-being: A framework for assessment. Washington, D.C.: Island Press. Millennium Ecosystem Assessment. (2005a). Ecosystems and human well-being: current stete and trends (Vol. 1). Washington, D.C.: Island Press. Millennium Ecosystem Assessment. (2005b). Ecosystems and Human Well-being: Synthesis. Washington, D.C.: Island Press. Millward, A. A., Torchia, M., Laursen, A. E., & Rothman, L. D. (2014). Vegetation Placement for Summer Built Surface Temperature Moderation in an Urban Microclimate. [journal article]. Environmental Management, 53(6), 1043-1057. Mitchell, R., & Popham, F. (2008). Effect of exposure to natural environment on health inequalities: an observational population study. The Lancet, 372(9650), 1655-1660. Moilanen, A., & Nieminen, M. (2002). Simple connectivity measures in spatial ecology. Ecology, 83(4), 1131-1145. Moles, R., Foley, W., Morrissey, J., & O'Regan, B. (2008). Practical appraisal of sustainable development—Methodologies for sustainability measurement at settlement level. Environmental Impact Assessment Review, 28(2), 144-165.

246

Monette, D. R., Sullivan, T. J., & DeJong, C. R. (2013). Applied social research: A tool for the human services: Cengage Learning. Monteith, J. (1965). Evaporation and the environment. 205-234. In the movement of water in living organisms, XIXth Symposium. Soc. of Exp. Biol., Swansea: Cambridge University Press. Moore, G. M. (2000). Treenet: A Management System and Choices for Australia. Paper presented at the TREENET Proceedings of the Inaugural Street Tree Symposium:7th and 8th September 2000, Adelaide. Moore, G. M. (2006). Urban trees and the global greenhouse. Paper presented at the TREENET Proceedings of the 7th National Street Tree Symposium: 7th and 8th September 2006, Adelaide. Myers, B., Cook, S., Maheepala, S., Pezzaniti, D., Beecham, S., et al. (2011). Interim water sensitive urban design targets for Greater Adelaide. Adelaide, Australia: Goyder Institute for Water Research. Nackaerts, K., Coppin, P., Muys, B., & Hermy, M. (2000). Sampling methodology for LAI measurements with LAI-2000 in small forest stands. Agricultural and Forest Meteorology, 101(4), 247-250. Nardo, M., Saisana, M., Saltelli, A., Tarantola, S., Hoffman, A., et al. (2005). Handbook on constructing composite indicators. Naumann, S., Davis, M., Kaphengst, T., Mav Pieterse, M., and Rayment, M. (2011). Design, implementation and cost elements of Green Infrastructure projects, Final report to the European Commission, DG Environment: Ecologic institute and GHK Consulting. NECR126 Natural England Commissioned. (2013). Green Infrastructure- Valuation Tools Assessment. London: Natural England. Nelson, E., Mendoza, G., Regetz, J., Polasky, S., Tallis, H., et al. (2009). Modeling multiple ecosystem services, biodiversity conservation, commodity production, and tradeoffs at landscape scales. Frontiers in Ecology and the Environment, 7(1), 4-11. Neuman, M. (2006). Infiltrating infrastructures: On the nature of networked infrastructure. Journal of Urban Technology, 13(1), 3-31. Neuman, M., & Smith, S. (2010). City planning and infrastructure: once and future partners. Journal of Planning History, 9(1), 21-42. Niemeijer, D., & de Groot, R. S. (2008). A conceptual framework for selecting environmental indicator sets. Ecological Indicators, 8(1), 14-25. Nieminen, T., Martelin, T., Koskinen, S., Aro, H., Alanen, E., et al. (2010). Social capital as a determinant of self-rated health and psychological well-being. International Journal of Public Health, 55(6), 531-542. North Carolina Division of Water Quality. (2007). Stormwater Best Management Practices Manual: NCDENR. Norton, B., Coutts, A., Livesley, S., & Williams, N. (2013). Technical Report: Decision principles for the selection and placement of green infrastructure. Australia: Victorian Centre for Climate Change Adaptation Research. Norton, B. A., Coutts, A. M., Livesley, S. J., Harris, R. J., Hunter, A. M., et al. (2015). Planning for cooler cities: A framework to prioritise green infrastructure to mitigate high temperatures in urban landscapes. Landscape and Urban Planning, 134, 127- 138. Noss, R. F., & Cooperrider, A. (1994). Saving nature's legacy: protecting and restoring biodiversity: Island Press. Noss, R. F., O'Connell, M., & Murphy, D. D. (1997). The science of conservation planning: habitat conservation under the Endangered Species Act: Island Press.

247

Nouri, H., Beecham, S., Kazemi, F., Hassanli, A., & Anderson, S. (2013). Remote sensing techniques for predicting evapotranspiration from mixed vegetated surfaces. Hydrology and Earth System Sciences Discussions, 10(3), 3897-3925. Nowak, D. J. (1994a). Air pollution removal by Chicago’s urban forest. Chicago’s urban forest ecosystem: Results of the Chicago urban forest climate project, 63-81. Nowak, D. J. (1994b). Atmospheric carbon dioxide reduction by Chicago’s urban forest. In E. G. McPherson, D. J. Nowak & R. A. Rowntree (Eds.), Chicago’s urban forest ecosystem: results of the Chicago Urban Forest Climate Project (pp. 83–94): US Department of Agriculture. Nowak, D. J. (1996). Notes: estimating leaf area and leaf biomass of open-grown deciduous urban trees. Forest science, 42(4), 504-507. Nowak, D. J. (2002). The effects of urban trees on air quality. USDA Forest Service, 96-102. Nowak, D. J., & Crane, D. E. (2000). The Urban Forest Effects (UFORE) Model: quantifying urban forest structure and functions. In M. Hansen & T. Burk (Eds.), Integrated tools for natural resources inventories in the 21st century. U.S. (pp. 714-720): Department of Agriculture, Forest ServiceNorth Central Forest Experiment Station, Boise. Nowak, D. J., & Crane, D. E. (2001). Carbon storage and sequestration by urban trees in the USA. Elsevier Science(116), 381-389. Nowak, D. J., & Crane, D. E. (2002). Carbon storage and sequestration by urban trees in the USA. Environmental Pollution, 116(3), 381-389. Nowak, D. J., Crane, D. E., & Stevens, J. C. (2006). Air pollution removal by urban trees and shrubs in the United States. Urban Forestry & Urban Greening, 4(3), 115-123. Nowak, D. J., Crane, D. E., Stevens, J. C., Hoehn, R. E., Walton, J. T., et al. (2008). A ground-based method of assessing urban forest structure and ecosystem services. Nowak, D. J., & Greenfield, E. J. (2012). Tree and impervious cover in the United States. Landscape and Urban Planning, 107(1), 21-30. Nowak, D. J., Greenfield, E. J., Hoehn, R. E., & Lapoint, E. (2013). Carbon storage and sequestration by trees in urban and community areas of the United States. Environmental Pollution, 178, 229-236. OECD. (1993). OECD core set of indicators for environmental performance reviews: A synthesis report by the group on the state of the environment. Paris. Oke, T. (1988). Street design and urban canopy layer climate. Energy and Buildings, 11(1), 103-113. Oke, T. R. (2002). Boundary layer climates: Routledge. Olmsted, Vaux, Co, Olmsted, & Vaux. (1868). Preliminary report upon the proposed suburban village at riverside, near Chicago: Sutton, Bowne. Onwuegbuzie, A. J., & Johnson, R. B. (2006). The validity issue in mixed research. Research in the Schools, 13(1), 48-63. Osmond, P. (2010). Hemispherical photography as a tool for urban sustainability evaluation and design. OIDA International Journal of Sustainable Development, 1(9), 63-74. Pakzad, P., & Osmond, P. (2015). A conceptual framework for assessing green infrastructure sustainability performance in Australia. Paper presented at the State of Australian Cities (SOAC) 2015. Pakzad, P., & Osmond, P. (2016). Developing a Sustainability Indicator Set for Measuring Green Infrastructure Performance. Procedia - Social and Behavioral Sciences, 216, 68-79. Pakzad, P., Osmond, P., & Corkery, L. (2017). Developing key sustainability indicators for assessing green infrastructure performance. Procedia Engineering, 180, 146-156.

248

Pakzad, P., Osmond, P., & Philipp, C. H. (2015). Review of tools for quantifying the contribution of green infrastructure to carbon performance. Paper presented at the ICUC9 - 9th International Conference on Urban Climate jointly with 12th Symposium on the Urban Environment, 20-24 July 2015, France. Panter, J., Jones, A., & Hillsdon, M. (2008). Equity of access to physical activity facilities in an English city. Preventive Medicine, 46(4), 303-307. Parker, K., Head, L., Chisholm, L. A., & Feneley, N. (2008). A conceptual model of ecological connectivity in the Shellharbour Local Government Area, New South Wales, Australia. Landscape and Urban Planning, 86(1), 47-59. Parramatta City Council. (2015). Life in our city: Parramatta biodiversity strategy 2015- 2025. Sydney, Australia. Pauleit, S., Liu, L., Ahern, J., & Kazmierczak, A. (2011). Multifunctional green infrastructure planning to promote ecological services in the city. Urban Ecology. Oxford University Press, Oxford, 272-286. Penwarden, A. D., & Wise, A. F. E. (1975). Wind environment around buildings: HMSO. Petty, N. J., Thomson, O. P., & Stew, G. (2012). Ready for a paradigm shift? Part 1: Introducing the philosophy of qualitative research. Manual Therapy, 17(4), 267-274. Pickett, S. T., Cadenasso, M. L., Grove, J. M., Nilon, C. H., Pouyat, R. V., et al. (2001). Urban ecological systems: Linking terrestrial ecological, physical, and socioeconomic components of metropolitan areas 1. Annual review of ecology and systematics, 32(1), 127-157. Pickup, J., & De Dear, R. (2000). An outdoor thermal comfort index (OUT_SET*)-part I-the model and its assumptions. Paper presented at the Biometeorology and urban climatology at the turn of the millenium. Selected Papers from the Conference ICB- ICUC. Pillsbury, N. H., Reimer, J. L., & Thompson, R. P. (1998). Tree volume equations for fifteen urban species in California: Tech. Rep. 7. San Luis Obispo, CA: Urban Forest Ecosystems Institute, California Polytechnic State University. Pitman, S., & Ely, M. (2013). From grey to green: Life support for human habitats. Paper presented at the 14 th National Street Tree Symposium 2013. Pitman, S. D., Daniels, C. B., & Ely, M. E. (2015). Green infrastructure as life support: urban nature and climate change. Transactions of the Royal Society of South Australia, 139(1), 97-112. Planet Ark. (2012). Planting Trees: Just What The Doctor Ordered. Australia: Planet Ark. Plant, L. (2006). Beautiful one day, perfect the next-Is there room for improvement? Paper presented at the TREENET Proceedings of the 7th National Street Tree Symposium: 7th and 8th September 2006, Adelaide. Pol, J. (2010). Design of an infrastructure sustainability assessment method for application in DGBC area. Pulcini, J., Wilbur, J., Allan, J., Hanson, C., & Uphold, C. R. (2006). Determining criteria for excellence in nurse practitioner education: Use of the Delphi Technique. Nursing outlook, 54(2), 102-110. Ranganathan, J., Bennett, K., Raudsepp-Hearne, C., Lucas, N., Irwin, F., et al. (2008). Ecosystem services: a guide for decision makers. world resources institute. Washington, D.C., 11(1). Rauscher, R. C., & Momtaz, S. (2016). Cities in Global Transition: Creating Sustainable Communities in Australia: Springer. Robitu, M., Musy, M., Inard, C., & Groleau, D. (2006). Modeling the influence of vegetation and water pond on urban microclimate. Solar Energy, 80(4), 435-447.

249

Román, M. O., Gatebe, C. K., Shuai, Y., Wang, Z., Gao, F., et al. (2013). Use of in situ and airborne multiangle data to assess MODIS-and Landsat-based estimates of directional reflectance and albedo. IEEE Transactions on Geoscience and Remote Sensing, 51(3), 1393-1404. Rowntree, R. A. (1984). Ecology of the urban forest—introduction to part I. Urban Ecology, 8(1-2), 1-11. Rückerl, R., Schneider, A., Breitner, S., Cyrys, J., & Peters, A. (2011). Health effects of particulate air pollution: a review of epidemiological evidence. Inhalation toxicology, 23(10), 555-592. Russo, A., J Escobedo, F., & Zerbe, S. (2016). Quantifying the local-scale ecosystem services provided by urban trees streetscapes in Bolzano, Italy. AIMS Environmental Science, 3(1), 58-76. Rutherford, J., Carter, M., & Christidis, K. (2013). Classification framework for public open space. Australasian Parks and Leisure, 16(1), 34. Sæbø, A., Popek, R., Nawrot, B., Hanslin, H., Gawronska, H., et al. (2012). Plant species differences in particulate matter accumulation on leaf surfaces. Science of the Total Environment, 427, 347-354. Sallis, J. F., Floyd, M. F., Rodríguez, D. A., & Saelens, B. E. (2012). Role of built environments in physical activity, obesity, and cardiovascular disease. Circulation, 125(5), 729-737. Sandstrom, P. M. (2002). Green nfrastructure planning in urban Sweden. Prac. Res. 17(4), 373-385. Sanesi, G., Colangelo, G., Lafortezza, R., Calvo, E., & Davies, C. (2017). Urban green infrastructure and urban forests: A case study of the Metropolitan Area of Milan. Landscape Research, 42(2), 164-175. Sanusi, R., Johnstone, D., May, P., & Livesley, S. J. (2016). Street orientation and side of the street greatly influence the microclimatic benefits street trees can provide in summer. Journal of environmental quality, 45(1), 167-174. Saunders, M., Lewis, P., & Thornhill, A. (2009). Research methods for business students (5 ed.). England: Prentice Hall. Sawka, M., Millward, A. A., Mckay, J., & Sarkovich, M. (2013). Growing summer energy conservation through residential tree planting. Landscape and Urban Planning, 113, 1-9. Schäffler, A., & Swilling, M. (2013). Valuing green infrastructure in an urban environment under pressure — The Johannesburg case. Ecological Economics(86), 246-257. Schueler, T. R. (1987). Controlling Urban Runoff: A Practical Manual for Planning and Designing Urban BMPsWashington Metropolitan Water Resources Planning Board. Washington, D.C. Schumaker, N. H. (1996). Using landscape indices to predict habitat connectivity. Ecology, 77(4), 1210-1225. Schwendenmann, L., & Mitchell, N. D. (2014). Carbon accumulation by native trees and soils in an urban park, Auckland. New Zealand Journal of Ecology, 213-220. Scott, C. J. (2002). Competencies needed by business teachers to work with students who have disabilities. Virginia Polytechnic Institute and State University. Segnestam, L. (2003). Indicators of environment and sustainable development: theories and practical experience (No. 89). World Bank: Washington, D.C. (EUA): Environment Department Environmental Economics

250

Seibert, M. J. (2004). The identification of strategic management counseling competencies essential for the Small Business and Technology Development Center: A modified Delphi study. Shashua‐Bar, L., Potchter, O., Bitan, A., Boltansky, D., & Yaakov, Y. (2010). Microclimate modelling of street tree species effects within the varied urban morphology in the Mediterranean city of Tel Aviv, Israel. International journal of climatology, 30(1), 44-57. Simon, D. (2016). Rethinking Sustainable Cities; accessible, green and fair: Bristol: Policy Press. SITES. (2009). SITES. The Sustainable Sites Initiative; Guidelines and Performanve Benchmarks. Retrieved from http://www.sustainablesites.org/report/Guidelines%20and%20Performance%20Bench marks_2009.pdf Slentz, C. A., Aiken, L. B., Houmard, J. A., Bales, C. W., Johnson, J. L., et al. (2005). Inactivity, exercise, and visceral fat. STRRIDE: a randomized, controlled study of exercise intensity and amount. Journal of Applied Physiology, 99(4), 1613-1618. Smith, F. W., Sampson, D. A., & Long, J. N. (1991). Notes: Comparison of leaf area index estimates from tree allometrics and measured light interception. Forest Science, 37(6), 1682-1688. Snyder, R., Pedras, C., Montazar, A., Henry, J., & Ackley, D. (2015). Advances in ET-based landscape irrigation management. Agricultural Water Management, 147, 187-197. Snyder, R. L., & Eching, S. (2005). Urban landscape evapotranspiration. Landscape water use. California water plan update, 4, 691-693. Spagnolo, J., & De Dear, R. (2003). A field study of thermal comfort in outdoor and semi- outdoor environments in subtropical Sydney Australia. Building and environment, 38(5), 721-738. Spagnolo, J., & De Dear, R. (2003). A human thermal climatology of subtropical Sydney. International journal of climatology, 23(11), 1383-1395. Spirn, A. W. (1985). Urban nature and human design. Bulletin ofthe German Historical Institute. Spitzer, M. A. (1999). Toward a sustainable America: Advancing prosperity, opportunity and a healthy environment for the 21st century: Diane publishing. Standard, A. (2010). Standard 55-2010:“Thermal Environmental Conditions for Human Occupancy”; ASHRAE. Atlanta USA. Stewart, I. D., & Oke, T. R. (2012). Local climate zones for urban temperature studies. Bulletin of the American Meteorological Society, 93(12), 1879-1900. Stigsdotter, U. K., Ekholm, O., Schipperijn, J., Toftager, M., Kamper-Jørgensen, F., et al. (2010). Health promoting outdoor environments-Associations between green space, and health, health-related quality of life and stress based on a Danish national representative survey. Scandinavian journal of public health. Sukhdev, P., Wittmer, H., Schröter-Schlaack, C., Nesshöver, C., Bishop, J., et al. (2010). The economics of ecosystems and biodiversity: mainstreaming the economics of nature: a synthesis of the approach, conclusions and recommendations of TEEB. European Communities. Taha, H. (1994). Aircraft-based albedo measurements over the South Coast Air Basin. Analysis of Energy Efficiency of Air Quality in the South Coast Air Basin-Phase II, Rep. No. LBL-35728, Lawrence Berkeley National Laboratory, Berkeley, CA, 43-59. Taha, H. (1997). Urban climates and heat islands: albedo, evapotranspiration, and anthropogenic heat. Energy and buildings, 25(2), 99-103.

251

Taha, H., Chang, S., & Akbari, H. (2000). Meteorological and air quality impacts of heat island mitigation measures in three US Cities. Lawrence Berkeley National Laboratory Report No. LBL-44222, Berkeley, CA. Takim, R., Akintoye, A., & Kelly, J. (2004). Analysis of measures of construction project success in Malaysia. Association of Researches in Construction Management, 2(9), 1123-1113. Talberth, J., Gray, E., Yonavjak, L., & Garther, T. (2013). Green versus gray: Nature’s solutions to infrastructure demands. The Solutions Journal, 4(1), 40-47. Tarran, J. (2006). Trees, Urban Ecology and Community Health. Paper presented at the TREENET Proceedings of the 7th National Street Tree Symposium: 7th and 8th September 2006, Adelaide. Tarran, J. (2009). People and trees: providing benefits, overcoming impediments. Paper presented at the TREENET Proceedings of the 10th National Street Tree Symposium 3rd and 4th September 2009. Taylor, P. D., Fahrig, L., Henein, K., & Merriam, G. (1993). Connectivity is a vital element of landscape structure. Oikos, 68(3), 571-573. Taylor, P. D., Fahrig, L., & With, K. A. (2006). Landscape connectivity: a return to the basics. In Kevin R. Crooks & M. Sanjayan (Eds.), Connectivity Conservation, Chapter: 2 (pp. 29-43): Cambridge TEP. (2005). Advancing the Delivery of green ifrastructure: Targeting Issues in England's Northwest.: The Environment Partnership. TEP. (2008). Towards a green infrastructure framework for Greater Manchester: Full report. UK: AGMA and Natural England. Teriman, S. (2012). Measuring neighbourhood sustainability: A comparative analysis of residential types in Malaysia. The NSW Government Planning and Environment. (2014). A plan for growing Sydney. The United States Department of Agriculture (Producer). (2006) i-Tree Applications. United States Department of Agriculture. retrieved from http://www.itreetools.org/index.php The US Environmental Protection Agency (EPA). (2016). Stormwater Planning. Community Solutions for Voluntary Long-Term Stormwater Planning Retrieved November, 2016, from https://www.epa.gov/npdes/stormwater-planning#technicalassistance Tigelaar, D. E., Dolmans, D. H., Wolfhagen, I. H., & Van der Vleuten, C. P. (2004). The development and validation of a framework for teaching competencies in higher education. Higher education, 48(2), 253-268. Tischendorf, L., & Fahrig, L. (2000). How should we measure landscape connectivity? Landscape ecology, 15(7), 633-641. Tiwary, A., Williams, I., Heidrich, O., Namdeo, A., Bandaru, V., et al. (2016). Development of multi-functional streetscape green infrastructure using a performance index approach. Environmental Pollution, 208, 209-220. Tongco, M. D. C. (2007). Purposive sampling as a tool for informant selection. Townend, M., & Weerasuriya, R. (2010). Beyond Blue to Green: The benefits of contact with nature for mental health and well-being, A Literature Review: School of Health and Social Development, Deakin University prepared for Beyond Blue: The National Depression Initiative. Tzoulas, K., Korpela, K., Venn, S., Yli-Pelkonen, V., Kaźmierczak, A., et al. (2007). Promoting ecosystem and human health in urban areas using Green Infrastructure: A literature review. Landscape and urban planning, 81(3), 167-178.

252

U.S. Environmental Protection Agency. (2015). Interagency Working Group on Social Cost of Carbon for Regulatory Impact Analysis Under Executive Order 12866: United States Government. U.S. Green Building Council. (2010). Guidance on Calculating Outdoor Water Demand Reduction: for LEED for Homes Credits SS 2.5 and/or WE 2.3. http://www.usgbc.org/Docs/Archive/General/Docs4943.pdf U.S. Green Building Council. (2015). LEED v2009 Water Use Reduction Calculator. http://www.usgbc.org/resources/2009-water-use-reduction-calculator Ulrich, R. (2002). Communicating with the healthcare community about plant benefits. Paper presented at the Proceedings of the Sixth International People Plant Symposium. Chicago: Chicago Botanic Garden. Ulrich, R. S., Simons, R. F., Losito, B. D., Fiorito, E., Miles, M. A., et al. (1991). Stress recovery during exposure to natural and urban environments. Journal of environmental psychology, 11(3), 201-230. United Nations. (2012). Population Division, World Urbanization Prospects: . NewYork: United Nations. United States Department of Agriculture Forest Service. (2016). i-Tree Eco v6. Retrieved October, 2016, from https://www.itreetools.org/eco/index.php Urban, D., & Keitt, T. (2001). Landscape connectivity: a graph‐theoretic perspective. Ecology, 82(5), 1205-1218. Van den Berg, A. E., Maas, J., Verheij, R. A., & Groenewegen, P. P. (2010). Green space as a buffer between stressful life events and health. Social science & medicine, 70(8), 1203-1210. Van der Ryn, S., & Cowan, S. (1996). Ecological design. Washington, D.C.: Island Press. Vandepol, J. F. (2010). Infrastructure Sustainability Assessment method Dutch: University of Twente. Vandermeulen, V., Verspecht, A., Vermeire, B., Van Huylenbroeck, G., & Gellynck, X. (2011). The use of economic valuation to create public support for green infrastructure investments in urban areas. Landscape and Urban Planning, 103(2), 198-206. Vojtech, E., Turnbull, L. A., & Hector, A. (2007). Differences in light interception in grass monocultures predict short-term competitive outcomes under productive conditions. PloS one, 2(6), e499. Volk, M., Hirschfeld, J., Dehnhardt, A., Schmidt, G., Bohn, C., et al. (2008). Integrated ecological-economic modelling of water pollution abatement management options in the Upper Ems River Basin. Ecological Economics, 66(1), 66-76. Vos, P. E., Maiheu, B., Vankerkom, J., & Janssen, S. (2013). Improving local air quality in cities: to tree or not to tree? Environmental pollution, 183, 113-122. Walker, B., Holling, C. S., Carpenter, S., & Kinzig, A. (2004). Resilience, adaptability and transformability in social–ecological systems. Ecology and society, 9(2). Wang, J., Endreny, T. A., & Nowak, D. J. (2008). Mechanistic Simulation of Tree Effects in an Urban Water Balance Model1: Wiley Online Library. Wang, Y., Bakker, F., de Groot, R., Wörtche, H., & Leemans, R. (2015). Effects of urban green infrastructure (UGI) on local outdoor microclimate during the growing season. Environmental Monitoring and Assessment, 187(12), 1-14. Wang, Z., Schaaf, C. B., Strahler, A. H., Chopping, M. J., Román, M. O., et al. (2014). Evaluation of MODIS albedo product (MCD43A) over grassland, agriculture and forest surface types during dormant and snow-covered periods. Remote Sensing of Environment, 140, 60-77.

253

Wania, A., Bruse, M., Blond, N., & Weber, C. (2012). Analysing the influence of different street vegetation on traffic-induced particle dispersion using microscale simulations. Journal of environmental management, 94(1), 91-101. Ward Thompson, C., Aspinall, P., Roe, J., Robertson, L., & Miller, D. (2016). Mitigating Stress and Supporting Health in Deprived Urban Communities: The Importance of Green Space and the Social Environment. International journal of environmental research and public health, 13(4), 440. Ward Thompson, C., & Aspinall, P. A. (2011). Natural environments and their impact on activity, health, and quality of life. Applied Psychology: Health and Well‐Being, 3(3), 230-260. Water Melbourne. (2016). MUSIC Guidelines; Recommended Input Parameters and Modelling Approaches for MUSIC Users in Melbourne Water’s service area. State Government of Victoria: Melbourne, Australia. Waters, J. (2016). Accessible cities: from urban density to multidimensional accessibility. In D. Simon (Ed.), Rethinking Sustainable Cities: Accessible, Green and Fair (pp. 11). UK: University of Bristol: Policy Press. Weber, T., Sloan, A., & Wolf, J. (2006). Maryland's Green Infrastructure Assessment: Development of a comprehensive approach to land conservation. Landscape and urban planning, 77(1), 94-110. Whitehead, T., Simmonds, D., & Preston, J. (2006). The effect of urban quality improvements on economic activity. Journal of Environmental Management, 80(1), 1-12. Willems, E., Vandevoort, C., Willekens, A., & Buffaria, B. (2000). Landscape and land cover diversity index From land cover to landscape diversity in the European Union: Report of the European Commission. Brussels, Belgium: Euroepan Commission. Wilson, J. W. (1963). Estimation of foliage denseness and foliage angle by inclined point quadrats. Australian journal of botany, 11(1), 95-105. Witten, K., Pearce, J., & Day, P. (2011). Neighbourhood Destination Accessibility Index: a GIS tool for measuring infrastructure support for neighbourhood physical activity. Environment and planning A, 43(1), 205-223. Wolch, J., Newell, J., Seymour, M., Huang, H. B., Reynolds, K., et al. (2010). The forgotten and the future: Reclaiming back alleys for a sustainable city. Environment and Planning A, 42(12), 2874-2896. Wolf, K. L. (2003). Ergonomics of the city: Green infrastructure and social benefits. Paper presented at the Engineering Green: Proceedings of the 11th National Urban Forest Conference Washington, D.C. Wong, T., Breen, P., & Lloyd, S. (2000). Water sensitive road design options for improving stormwater quality of road runoff: Technical Report for Cooperative Research Centre for Catchment Hydrology 00/1. Wong, T. H. F., Allen, R., Beringer, J., Brown, R. R., Deletić, A., et al. (2012). Blueprint 2012- Stormwater Management in a Water Sensitive City. Melbourne, Australia: The Centre for Water Sensitive Cities, Monash University. World Commission on Environment and Development. (1989). Our common future. By World commission on environment and development.(London, Oxford University Press, 1987, pp. 383£ 5.95.): Wiley Online Library. World Health Organisation (WHO). (2016). Urban green spaces and health; a review of evodence. Copenhagen: WHO Regional Office for Europe. World Health Organization. (2011). Health economic assessment tools (HEAT) for walking and for cycling. Copenhagen: WHO Regional Office for Europe.

254

World Health Organization (WHO). (1946). Preamble to the constitution of th World Health Organization as adopted by the International Health Conference, New York, 1946. Entered into force on 7 April 1948. New York, 19-22. Wright, J. L. (1982). New evapotranspiration crop coefficients. Proceedings of the American Society of Civil Engineers, Journal of the Irrigation and Drainage Division, 108(IR2), 57-74. Wright, J. P., Flecker, A. S., & Jones, C. G. (2003). Local vs. landscape controls on plant species richness in beaver meadows. Ecology, 84(12), 3162-3173. Wu, J., & Hobbs, R. J. (2007). Key topics in landscape ecology: Cambridge University Press. Yang, J., McBride, J., Zhou, J., & Sun, Z. (2005). The urban forest in Beijing and its role in air pollution reduction. Urban Forestry & Urban Greening, 3(2), 65-78. Yang, J., Yu, Q., & Gong, P. (2008). Quantifying air pollution removal by green roofs in Chicago. Atmospheric environment, 42(31), 7266-7273. Yli-Pelkonen, V. (2008). Ecological information in the political decision making of urban land-use planning. Journal of environmental planning and management, 51(3), 345- 362. Young, R. F. (2011). Planting the living city: Best practices in planning green infrastructure—Results from major us cities. Journal of the American Planning Association, 77(4), 368-381. Youngquist, T. D. (2009). What is green infrastructure? An evaluation of green infrastructure plans from across the United States. Iowa State University, Ames, lowa. Yu, S., Yu, B., Song, W., Wu, B., Zhou, J., et al. (2016). View-based greenery: A three- dimensional assessment of city buildings’ green visibility using floor green view index. Landscape and Urban Planning, 152, 13-26. Zeller, K. A., McGarigal, K., & Whiteley, A. R. (2012). Estimating landscape resistance to movement: a review. Landscape Ecology, 27(6), 777-797. Zhao, M., Kong, Z.-h., Escobedo, F. J., & Gao, J. (2010). Impacts of urban forests on offsetting carbon emissions from industrial energy use in Hangzhou, China. Journal of Environmental Management, 91(4), 807-813. Zhong, Q., & Yinhai, L. (1988). Satellite observation of surface albedo over the Qinghai- Xizang plateau region. Advances in atmospheric sciences, 5(1), 57-65.

255

256

APPENDIX A: Online questionnaire form

Table A. 1 Blank questionnaire form

SECTION A: Verify green infrastructure performance framework Table B. 1 Online questionnaire form

Q1: Are you familiar with the term “Green infrastructure”? o Yes o No

Q2: Are you familiar with the term “Triple Bottom-Line” of the sustainability concept? o Yes o No

Q3: Are you familiar with concept of “Millennium Ecosystem Assessment” framework? o Yes o No

Q4: These definitions of GI have been derived from the results of semi-structured interviews. Please indicate the extent to which you agree/disagree with each of the following definitions of “Green infrastructure”. Neither Definitions Disagree agree or Agree disagree Definition 1: Green infrastructure (GI) is a policy and strategic approach to land and species conservation. o o o

Definition 2: GI is a network of energy, materials and species flows that maintains and improves ecological functions in combination with multifunctional land uses and provides associated benefits to human o o o populations and ecosystems.

Definition 3: GI refers to the integration of ecological functions through natural and engineered networks into conventional infrastructure systems to enhance their functions, and it can o o o significantly reduce their carbon footprint.

Definition 4: GI is an ecological solution underpinned by the concept of ecosystem services to improve the sustainability level of the urban and built environment. It embraces the idea of the triple bottom line – o o o the social, economic and environmental aspects of the urban environment.

o If none of these definitions please specify your own definition:

Q5: Three frameworks for measuring the sustainability performance of green infrastructure have been nominated by experts interviewed by this researcher. Please indicate the extent to which you agree/disagree with each of the possible frameworks set out below. Neither agree or Possible frameworks for GI Disagree Agree disagree TBL sustainability framework (Environmental, Social and Economic) can be an appropriate framework for o o o measuring GI performance The Millennium Ecosystem Assessment framework (Provisioning services, Regulating services, Cultural o o o services, Supporting services + Biodiversity) can be an appropriate framework for measuring GI performance 257

A combination of both frameworks (TBL and MEA) can be an appropriate framework for measuring GI o o o performance o If none of these frameworks please specify your own alternatives:

Q6: Please rate the importance to you of each of the categories of ecosystem services set out below, from 0 = Less important to 4 = most important. Categories 0 1 2 3 4 Provisioning services (e.g. food, water, o o o o o timber, genetic resources) Regulating services (e.g. Climate regulation, air and water quality, carbon o o o o o sequestration) Cultural/Social services (e.g. recreational, o o o o o aesthetic and spiritual) Supporting/habitat services (e.g. soil o o o o o formation, nutrient cycling, photosynthesis)

Q7: Please rate the degree of the importance of each aspect of sustainability from 0 = Less important to 4 = most important.

Categories 0 1 2 3 4 Environmental o o o o o Social (including Cultural) o o o o o Economic o o o o o

SECTION B: Rating indicators

The list of indicators below has been established based on literature review and interview results. Q8: These categories below are derived from the combination of both frameworks (TBL and MEA) Please assign a weight to each category in terms of their importance in the index (Sum of all weight should be 100%).

WEIGHT OF EACH CATEGORY CATEGORIES (0%-100%) Ecological indicators Human health indicators Socio-cultural indicators Economic indicators Total= 100%

Q9: Ecological indicators: How do you rate each of the following possible performance of GI, from 0 = Less important to 4 = most important. ECOLOGICAL INDICATORS 0 1 2 3 4 Climate and microclimatic modifications (e.g. Urban Heat Island effect mitigation; temperature moderation through o o o o o evapotranspiration and shading; wind speed modification) Air quality improvement (e.g. Pollutant o o o o o removal; Avoided emissions) 258

Carbon Emissions (e.g. direct carbon sequestration and storage; o o o o o avoided greenhouse gas emissions through cooling) Reduced building energy use for heating and cooling (through e.g. shading by trees; o o o o o covering building by green roof and green walls) Hydrological regulation (e.g. flow control and flood reduction; o o o o o regulation of water quality; water purification) Improved soil quality and Erosion prevention (e.g. soil fertility; soil o o o o o stabilisation) Waste decomposition and nutrient cycling o o o o o Noise level attenuation o o o o o Biodiversity-protection and enhancement (e.g. Communities; species; genetic o o o o o resources; habitats)

Q10: Do you know any other indicator(s) that can be included in the list of ecological indicators? o Yes o No If yes, Please specify your proposed indicator(s) and assign a rate from 0 to 4:

Q11: Health and well-being indicators: How do you rate each of the following possible performance of GI, from 0 = Less important to 4 = most important. HEALTH INDICATORS 0 1 2 3 4 Improving physical well-being ( e.g. physical outdoor activity; healthy food; o o o o o healthy environments ) Improving social well-being (e.g. social interaction; social integration; community o o o o o cohesion) Improving mental well-being (e.g. reduced depression and anxiety; recovery o o o o o from stress; attention restoration; positive emotions) Others (Please specify): o o o o o

Q12: Do you know any other indicator(s) that can be included in the list of health and well- being indicators? o Yes o No If yes, Please specify your proposed indicator(s) and assign a rate from 0 to 4:

Q13: Socio-cultural indicators: How do you rate each of the following possible performance of GI, from 0 = Less important to 4 = most important.

SOCIO-CULTURAL INDICATORS 0 1 2 3 4 Food production (e.g. urban agriculture; kitchen gardens; edible landscape and o o o o o community gardens) Opportunities for recreation, tourism and social interaction (community o o o o o livability) Improving pedestrian ways and their o o o o o connectivity

259

(e.g. increasing safety; quality of path; connectivity and linkage with other modes) Improving accessibility o o o o o Provision of outdoor sites for education o o o o o and research Reduction of crimes and fear of crime o o o o o (comfort; amenity and safety) Attachment to place and sense of o o o o o belonging (cultural and symbolic value) Enhancing attractiveness of cities (e.g. enhancing desirable views; restricting o o o o o undesirable views )

Q14: Do you know any other indicator(s) that can be included in the list of socio-cultural indicators? o Yes o No If yes, Please specify your proposed indicator(s) and assign a rate from 0 to 4:

Q15: Economic indicators: How do you rate each of the following possible performance of GI, from 0 = Less important to 4 = most important. ECONOMIC INDICATORS 0 1 2 3 4 Increased property values o o o o o Greater local economic activity (e.g. o o o o o tourism, recreation, cultural activities) Healthcare cost savings o o o o o Economic benefits of provision services (e.g. raw materials; timber; o o o o o food products; biofuels; medicinal products; fresh water etc.) Value of avoided CO2 emissions and o o o o o carbon sequestration Value of avoided energy consumption (e.g. reduced demands for cooling and o o o o o heating) Value of air pollutant o o o o o removal/avoidance Value of avoided grey infrastructure design(construction and management o o o o o costs) Value of reduced flood damage o o o o o Reducing cost of using private car by increasing walking and cycling (e.g. o o o o o shifting travel mode)

Q16: Do you know any other indicator(s) that can be included in the list of economic indicators? o Yes o No If yes, Please specify your proposed indicator(s) and assign a rate from 0 to 4:

SECTION C: Expert classification Q17: Please identify the job sector you are employed in? o Practitioner o Academic o Government

260

Q18: Please identify your industry sector(s)? o Building/Architecture o Infrastructure o Landscape o Urban planning

o Arboriculture/horticult o Ecology o Economy o Other (Please specify) ure

Q19: Please nominate your professional expertise? o Design and planning o Engineering o Policy making o Teaching o Research o Modelling o Others (Please ...... specify)

Q20: Please identify the main area(s) you focus on in your work (select all that apply) o Materials o Water o Energy o Carbon o Soil o Ecology o Waste o Noise o Climate change o Health and well- being o Transport o Heritage o Safety o Others (Please ...... specify)

Q21: How long have you been in this field? o Less than 2 o 2-5 years o 5-10 years o 10-20 years o More than 20 years years

Q22: Is your current location in Australia? o Yes o No If No, Please Specify your location:

Q23: If yes, What is your current location in Australia? o New South o Queensland o South o Tasmania o Victoria o Western Wales Australia Australia

Q24: Please give details of any of your projects (National and/or international), which are relevant to this study (project name, location, and year):

Q25: Have you published academic/research papers which are relevant to this study? o Yes o No If yes, Please give details of some of them that can be helpful for this study (title/year of publication):

Q26: Have you been involved in developing any assessment tools, models or frameworks? o Yes o No If yes, Please give details(Name the tool(s)/your responsibility):

261

APPENDIX B: Summary of questionnaire results

Table B. 1 Descriptive statistic of answers for questions 1 to 3 in section A Q2: Are you familiar with Q3: Are you familiar with Q1: Are you familiar with the term "Triple-Bottom- concept of "Millennium Statistics the term "Green Line" of the sustainability Ecosystem Assessment" Infrastructure"? concept? framework? Valid 373 373 373 N Missing 0 0 0 Mean 1.0375 1.2306 1.6488 Std. Error of Mean .00985 .02184 .02475 Median 1.0375a 1.2306a 1.6488a Mode 1.00 1.00 2.00 Std. Deviation .19032 .42176 .47799 Variance .036 .178 .228 Skewness 4.886 1.285 -.626 Std. Error of Skewness .126 .126 .126 Kurtosis 21.992 -.352 -1.617 Std. Error of Kurtosis .252 .252 .252 Range 1.00 1.00 1.00 Minimum 1.00 1.00 1.00 Maximum 2.00 2.00 2.00 a. Calculated from grouped data.

Table B. 2 Descriptive statistic of answers for questions 4 in section A Statistics Definition 1 Definition 2 Definition 3 Definition 4 Valid 373 373 373 373 N Missing 0 0 0 0 Mean 1.8928 2.4504 2.6005 2.7131 Std. Error of Mean .04242 .03483 .03238 .02845 Median 1.8496a 2.5015a 2.6493a 2.7493a Mode 1.00 3.00 3.00 3.00 Std. Deviation .81931 .67269 .62536 .54945 Variance .671 .453 .391 .302 Skewness .200 -.830 -1.314 -1.785 Std. Error of Skewness .126 .126 .126 .126 Kurtosis -1.483 -.459 .589 2.231 Std. Error of Kurtosis .252 .252 .252 .252 Range 2.00 2.00 2.00 2.00 Minimum 1.00 1.00 1.00 1.00 262

Maximum 3.00 3.00 3.00 3.00 a. Calculated from grouped data.

Table B. 3 Descriptive statistic of answers for questions 5 in section A F3: Combination of TBL & Statistics F1:TBL framework F2: MEA framework MEA Valid 373 373 373 N Missing 0 0 0 Mean 2.5201 2.3351 2.5603 Std. Error of Mean .03166 .03027 .02840 Median 2.5543a 2.3561a 2.5758a Mode 3.00 2.00 3.00 Std. Deviation .61149 .58455 .54844 Variance .374 .342 .301 Skewness -.893 -.228 -.732 Std. Error of Skewness .126 .126 .126 Kurtosis -.210 -.656 -.559 Std. Error of Kurtosis .252 .252 .252 Range 2.00 2.00 2.00 Minimum 1.00 1.00 1.00 Maximum 3.00 3.00 3.00 a. Calculated from grouped data.

Case Processing Summary N % Cases Valid 373 100.0 Excludeda 0 .0 Total 373 100.0 a. Listwise deletion based on all variables in the procedure.

Reliability Statistics Cronbach's Alpha Based on Standardized Cronbach's Alpha Items N of Items .929 .930 30

263

Item Statistics Indicators Mean Std. Deviation N Climate and microclimatic modifications (e.g. Urban Heat Island effect mitigation; temperature moderation through evapotranspiration and 4.3887 .75959 373 shading;wind speed modification) Air quality improvement (e.g. Pollutant removal; Avoided emissions) 4.3056 .74624 373 Carbon Emissions (e.g. direct carbon sequestration and storage; avoided greenhouse 4.2681 .86009 373 gas emissions through cooling) Reduced building energy use for heating and cooling (through e.g. shading by trees; covering 4.2761 .79406 373 building by green roof and green walls) Hydrological regulation (e.g. flow control and flood reduction; regulation of water quality; water 4.3110 .77238 373 purification) Improved soil quality and Erosion prevention (e.g. soil fertility; soil stabilisation) 3.9383 .89139 373 Waste decomposition and nutrient cycling 3.7989 .92736 373 Noise level attenuation 3.2279 .96111 373 Biodiversity-protection and enhancement (e.g. Communities; species; genetic resources; 4.2198 .88896 373 habitats) Improving physical well-being ( e.g. physical outdoor activity; healthy food; healthy 4.5335 .69335 373 environments ) Improving social well-being (e.g. social interaction; social integration; community 4.2064 .84099 373 cohesion) Improving mental well-being (e.g. reduced depression and anxiety; recovery from stress; 4.3298 .80060 373 attention restoration; positive emotions) Food production (e.g. urban agriculture; kitchen gardens; edible landscape and community 4.0349 .92537 373 gardens) Opportunities for recreation, tourism and social interaction (community livability) 4.0107 .80649 373 Improving pedestrian ways and their connectivity (e.g. increasing safety; quality of 4.1903 .86678 373 path; connectivity and linkage with other modes) Improving accessibility 3.9786 .91262 373 Provision of outdoor sites for education and research 3.6381 .96182 373 Reduction of crimes and fear of crime (comfort; amenity and safety) 3.6327 1.01425 373 Attachment to place and sense of belonging (cultural and symbolic value) 3.9598 .93094 373 Enhancing attractiveness of cities (e.g. enhancing desirable views; restricting undesirable 3.8633 .93617 373 views) Increased property values 3.2627 1.13859 373 Greater local economic activity (e.g. tourism, recreation, cultural activities) 3.8525 .95446 373 Healthcare cost savings 3.9517 .98801 373 Economic benefits of provision services (e.g. raw materials; timber; food products; biofuels; 3.7828 .99110 373 medicinal products; fresh water etc.) Value of avoided CO2 emissions and carbon sequestration 4.0322 .97773 373

264

Value of avoided energy consumption (e.g. reduced demands for cooling and heating) 4.2761 .84649 373 Value of air pollutant removal/avoidance 4.1046 .87517 373 Value of avoided grey infrastructure design(construction and management costs) 3.9437 .95295 373 Value of reduced flood damage 3.9598 .96497 373 Reducing cost of using private car by increasing walking and cycling (e.g. shifting 4.0080 .96578 373 travel mode)

Summary Item Statistics Maximum / Mean Minimum Maximum Range Minimum Variance N of Items

Item Means 4.010 3.228 4.534 1.306 1.404 .091 30 Item Variances .815 .481 1.296 .816 2.697 .029 30 Inter-Item .307 -.023 .710 .733 -31.294 .009 30 Correlations

Scale Statistics Mean Variance Std. Deviation N of Items 120.2869 239.969 15.49092 30

Intraclass Correlation Coefficient

Intraclass 95% Confidence Interval F Test with True Value 0 Correlationb Lower Bound Upper Bound Value df1 df2 Sig Single Measures .304a .272 .339 14.090 372 10788 .000 Average Measures .929c .918 .939 14.090 372 10788 .000 Two-way mixed effects model where people effects are random and measures effects are fixed. a. The estimator is the same, whether the interaction effect is present or not. b. Type C intraclass correlation coefficients using a consistency definition. The between-measure variance is excluded from the denominator variance. c. This estimate is computed assuming the interaction effect is absent, because it is not estimable otherwise.

265

Table B. 4 Weighted average index (WAI)- Ecological indicators ECOLOGICAL INDICATORS CATEGORY Score Weight 1 2 3 4 5 weight in Ecological Indicators WAI each (0.2) (0.4) (0.6) (0.8) (1) sub- indicator category

Climate and microclimatic modifications (e.g. Urban Heat Island effect mitigation; temperature moderation 2 4 39 130 198 0.878 11.95 3.874 through evapotranspiration and shading;wind speed modification)

Air quality improvement (e.g. Pollutant 0 5 49 146 173 0.861 11.72 3.801 removal; Avoided emissions) Carbon Emissions (e.g. direct carbon sequestration and storage; avoided 2 10 59 117 185 0.854 11.62 3.768 greenhouse gas emissions through cooling) Reduced building energy use for heating and cooling (through e.g. 0 12 44 146 171 0.855 11.64 3.775 shading by trees; covering building by green roof and green walls) Hydrological regulation (e.g. flow control and flood reduction; regulation of 2 6 41 149 175 0.862 11.74 3.806 water quality; water purification) Improved soil quality and Erosion prevention (e.g. soil fertility; soil 2 20 88 152 111 0.788 10.72 3.477 stabilisation) Waste decomposition and nutrient 3 27 106 143 94 0.760 10.34 3.354 cycling Noise level attenuation 10 72 151 103 37 0.646 8.79 2.850 Biodiversity-protection and enhancement (e.g. Communities; 1 16 61 117 178 0.844 11.49 3.725 species; genetic resources; habitats) Sum 373 7.347 100 32.43

Table B. 5 Weighted average index (WAI)- Health indicators HEALTH INDICATORS CATEGORY Score Weight 1 2 3 4 5 weight in Health Indicators WAI each (0.2) (0.4) (0.6) (0.8) (1) sub- indicator category Improving physical well-being ( e.g. physical outdoor activity; healthy food; 2 4 19 116 232 0.907 34.69 8.987 healthy environments ) Improving social well-being (e.g. social interaction; social integration; 2 14 47 152 158 0.841 32.18 8.339 community cohesion) Improving mental well-being (e.g. reduced depression and anxiety; recovery 3 8 36 142 184 0.866 33.13 8.584 from stress; attention restoration; positive emotions) Sum 373 2.614 100.00 25.910

266

Table B. 6 Weighted average index (WAI)- Sociocultural indicators SOCIO-CULTURAL CATEGORY Score weight Weight 1 2 3 4 5 Socio-Cultural Indicators WAI in sub- each (0.2) (0.4) (0.6) (0.8) (1) categor indicator y Food production (e.g. urban agriculture; kitchen gardens; edible 3 20 75 138 137 0.807 12.89 2.688 landscape and community gardens) Opportunities for recreation, tourism and social interaction (community 2 11 74 180 106 0.802 12.81 2.672 livability) Improving pedestrian ways and their connectivity (e.g. increasing safety; 4 10 57 142 160 0.838 13.38 2.792 quality of path; connectivity and linkage with other modes) Improving accessibility 6 16 75 159 117 0.796 12.71 2.651 Provision of outdoor sites for 5 39 117 137 75 0.728 11.62 2.424 education and research Reduction of crimes and fear of crime 10 42 99 146 76 0.727 11.60 2.420 (comfort; amenity and safety) Attachment to place and sense of 3 26 73 152 119 0.792 12.65 2.638 belonging (cultural and symbolic value) Enhancing attractiveness of cities (e.g. enhancing desirable views; restricting 4 26 90 150 103 0.773 12.34 2.574 undesirable views) Sum 373 6.262 100.00 20.860

Table B. 7 Weighted average index (WAI)- Economic indicators ECONOMIC CATEGORY Score Weight 1 2 3 4 5 weight in Economic Indicators WAI each (0.2) (0.4) (0.6) (0.8) (1) sub- indicator category Increased property values 28 66 114 110 55 0.65 8.33 1.732 Greater local economic activity (e.g. 7 24 87 154 101 0.77 9.83 2.046 tourism, recreation, cultural activities) Healthcare cost savings 6 27 74 138 128 0.79 10.09 2.098 Economic benefits of provision services (e.g. raw materials; timber; food 10 24 100 142 97 0.76 9.66 2.009 products; biofuels; medicinal products; fresh water etc.) Value of avoided CO2 emissions and 6 25 61 140 141 0.81 10.29 2.141 carbon sequestration Value of avoided energy consumption (e.g. reduced demands for cooling and 5 10 36 148 174 0.86 10.92 2.270 heating) Value of air pollutant 5 9 68 151 140 0.82 10.48 2.179 removal/avoidance Value of avoided grey infrastructure design (construction and management 6 18 90 136 123 0.79 10.07 2.094 costs) Value of reduced flood damage 5 28 67 150 123 0.79 10.11 2.102 Reducing cost of using private car by increasing walking and cycling (e.g. 6 21 73 137 136 0.80 10.23 2.128 shifting travel mode) Sum 373 7.83 100.00 20.80

267

APPENDIX C: Frequency of species and calculation of Shannon diversity score

Shannon Diversity Score = 2.9/ ln (53) = 0.73

Table C. 1 Species diversity

Species Name Frequency pi lnpi pi*lnpi Sum * -1 Acacia spp 5 0.002809 -5.87493 -0.0165 2.9 Angophora spp 1 0.000562 -7.48437 -0.0042 Araucaria spp 3 0.001685 -6.38576 -0.01076 Ash spp 31 0.017416 -4.05038 -0.07054 Banksia spp 2 0.001124 -6.79122 -0.00763 Bottlebrush spp 178 0.1 -2.30259 -0.23026 Brachychiton spp 27 0.015169 -4.18853 -0.06353 Calodendrum spp 3 0.001685 -6.38576 -0.01076 Cedar spp 1 0.000562 -7.48437 -0.0042 Cheesewood spp 2 0.001124 -6.79122 -0.00763 Cinnamon spp 13 0.007303 -4.91942 -0.03593 Corymbia spp 79 0.044382 -3.11492 -0.13825 Cottonwood spp 3 0.001685 -6.38576 -0.01076 Cypress spp 1 0.000562 -7.48437 -0.0042 Date palm spp 31 0.017416 -4.05038 -0.07054 Elaeocarpus spp 20 0.011236 -4.48864 -0.05043 Elm spp 4 0.002247 -6.09807 -0.0137 Fig spp 9 0.005056 -5.28714 -0.02673 Flindersia spp 128 0.07191 -2.63234 -0.18929 Grevillea spp 4 0.002247 -6.09807 -0.0137 Gum spp 61 0.03427 -3.37349 -0.11561 Hackberry spp 10 0.005618 -5.18178 -0.02911 Jacaranda spp 150 0.08427 -2.47373 -0.20846 Koelreuteria spp 7 0.003933 -5.53846 -0.02178 Lagerstroemia spp 12 0.006742 -4.99946 -0.0337 Lagunaria spp 1 0.000562 -7.48437 -0.0042 Locust spp 7 0.003933 -5.53846 -0.02178 Lophostemon spp 433 0.243258 -1.41363 -0.34388 Magnolia spp 8 0.004494 -5.40493 -0.02429 Maple spp 17 0.009551 -4.65116 -0.04442 Melaleuca spp 57 0.032022 -3.44132 -0.1102 Melia spp 3 0.001685 -6.38576 -0.01076 Michelia spp 4 0.002247 -6.09807 -0.0137 Oak spp 2 0.001124 -6.79122 -0.00763 Oleander spp 2 0.001124 -6.79122 -0.00763

268

Olive spp 7 0.003933 -5.53846 -0.02178 Pear spp 71 0.039888 -3.22169 -0.12851 Peppertree spp 8 0.004494 -5.40493 -0.02429 Pine spp 4 0.002247 -6.09807 -0.0137 Pistache spp 18 0.010112 -4.594 -0.04646 Plum spp 1 0.000562 -7.48437 -0.0042 Privet spp 2 0.001124 -6.79122 -0.00763 Robinia spp 11 0.00618 -5.08647 -0.03143 Sheoak spp 13 0.007303 -4.91942 -0.03593 Stenocarpus spp 17 0.009551 -4.65116 -0.04442 Sweetgum spp 61 0.03427 -3.37349 -0.11561 Syagrus spp 1 0.000562 -7.48437 -0.0042 Sycamore spp 190 0.106742 -2.23734 -0.23882 Tallowtree spp 17 0.009551 -4.65116 -0.04442 Tristaniopsis spp 19 0.010674 -4.53993 -0.04846 Tupelo spp 2 0.001124 -6.79122 -0.00763 Turpentine tree spp 1 0.000562 -7.48437 -0.0042 Washingtonia spp 18 0.010112 -4.594 -0.04646

269