Evaluating the Implementation of Compact Activity Centres in Greater Brisbane

Evaluating the implementation of compact activity centres in greater Brisbane

Mark Limb

Masters of Urban and Regional Planning

Submitted in fulfilment of the requirements for the degree of Doctor of Philosophy

School of Civil Engineering and Built Environment

Science and Engineering Faculty

Queensland University of Technology

2019

Table of Contents

Contents

1. INTRODUCTION ...... 16

1.1. BACKGROUND ...... 16

1.2. RESEARCH QUESTIONS ...... 18

1.3. RESEARCH APPROACH ...... 19

1.4. RESEARCH OUTCOMES ...... 22

1.5. DISSERTATION OUTLINE ...... 22

2. THE COMPACT CITY, ACTIVITY CENTRES, AND THE EVOLUTION OF PLANNING FOR GREATER BRISBANE ...... 25

2.1. THE RATIONALE AND COMMON CRITIQUES OF THE COMPACT CITY ...... 25

2.2. PLANNING FOR COMPACT ACTIVITY CENTRES IN BRISBANE ...... 31

2.3. CONCLUSION ...... 48

3. PLAN EVALUATION THEORY, RESULTS, AND METHODS ...... 49

3.1. PLAN EVALUATION ...... 50

3.2. EVALUATION OF THE IMPLEMENTATION OF COMPACT ACTIVITY CENTRES ...... 54

3.3. METHODS OF PLAN EVALUATION ...... 58

3.4. CONCLUSION ...... 67

4. RESEARCH PLAN ...... 69

4.1. RESEARCH QUESTIONS ...... 69

4.2. RESEARCH METHOD ...... 71

5. POLICY INTENT AND REALITY: CONFORMANCE OF 20 YEARS OF METROPOLITAN COMPACT ACTIVITY CENTRE POLICY IN GREATER BRISBANE ...... 125

5.1. CONFORMANCE WITH OBJECTIVES FOR HIGHER RESIDENTIAL DENSITIES ...... 126

5.2. CONFORMANCE WITH OBJECTIVES FOR A DIVERSITY OF DWELLING TYPES...... 134

5.3. CONFORMANCE WITH OBJECTIVES FOR EMPLOYMENT ...... 142

5.4. CONFORMANCE WITH OBJECTIVES FOR A MIXTURE OF USES AND ACCESSIBILITY ...... 152

5.5. DISCUSSION OF RESULTS ...... 162

5.6. CONCLUSION ...... 182

6. PERFORMANCE OF POLICY IN INFLUENCING THE DEVELOPMENT AND APPLICATION OF LAND USE REGULATIONS ...... 184

6.1. CONTENT ANALYSIS OF THE PERFORMANCE OF LOCAL GOVERNMENT LAND USE REGULATIONS ...... 185

iii

6.2. CHANGES TO LOCAL GOVERNMENT LAND USE REGULATIONS AND THEIR ALIGNMENT TO ACTIVITY CENTRE POLICY ...... 188

6.3. CONFORMANCE OF LAND USE CHANGES TO PLANNING REGULATIONS ...... 202

6.4. DISCUSSION AND CONCLUSION ...... 213

7. COMPACT ACTIVITY CENTRE PROGRESS AND ITS RELATIONSHIP TO POSSIBLE EXPLANATORY FACTORS ...... 216

7.1. RELATIONSHIPS BETWEEN COMPACT CENTRE INTENSIFICATION AND FACTORS ...... 217

7.2. RELATIONSHIPS WITH COMPACT CENTRE INTENSIFICATION WHILE CONTROLLING FOR THE INFLUENCE OF OTHER VARIABLES ...... 221

7.3. DISCUSSION AND CONCLUSION ...... 226

8. CONCLUSION AND DISCUSSION OF RESULTS ...... 229

8.1. OVERVIEW OF RESEARCH RESULTS ...... 230

8.2. RESEARCH IMPLICATIONS ...... 233

8.3. DIRECTIONS FOR FUTURE RESEARCH ...... 246

8.4. A FINAL SUMMARY ...... 253

9. REFERENCES ...... 256

10. APPENDICES ...... 279

10.1. APPENDIX 1 – INPUT DATA AND MANUAL ADJUSTMENT OF WALKABLE CATCHMENT DATA ...... 279

10.2. APPENDIX 2 – COMPACTNESS INDICATOR DATA TABLES ...... 283

10.3. APPENDIX 3 – ADJUSTMENTS TO SHOPPING CENTRE AREAS ...... 297

10.4. APPENDIX 4 – DETAILS OF DOCUMENTS ANALYSED FOR PLAN PERFORMANCE ...... 299

10.5. APPENDIX 5 – SUMMED LAND AREA OF DIS SCORES BY CENTRE AND YEAR ...... 316

10.6. APPENDIX 6 – CLASSIFICATION CRITERIA FOR LAND USE CONFORMANCE ...... 322

10.7. APPENDIX 7 - CORRELATION MATRIX OF INDEPENDENT VARIABLES ...... 324

10.8. APPENDIX 7 – SPRINGWOOD SUMMIT EVENT FLYER ...... 325

Table of Figures

FIGURE 1 - COMPOSITION OF THE REGIONAL PLANNING ADVISORY GROUP (ABBOTT 1995) ...... 33

FIGURE 2 - PROPOSED CENTRE NETWORK FROM THE 1993 POLICY PAPERS (THE STATE OF QUEENSLAND, 1993B) ...... 35

FIGURE 3 - "THE PREFERRED PATTERN" (THE STATE OF QUEENSLAND, 1993A) ...... 37

FIGURE 4 - A MORE "FLEXIBLE" IMPLEMENTATION PROCESS (THE STATE OF QUEENSLAND, 1994A) ...... 38

FIGURE 5 - MAJOR DISTRICT CENTRE SCOPING STUDY CANDIDATE CENTRES (PLANNING WORKSHOP AUSTRALIA, 1996)...... 39

FIGURE 6 - 2005 ACTIVITY CENTRE NETWORK (THE STATE OF QUEENSLAND 2005C) ...... 42

FIGURE 7 - RESIDENTIAL DENSITIES FOR ACTIVITY CENTRES IN THE SEQRP2017 (THE STATE OF QUEENSLAND 2016B) ...... 47

FIGURE 8 – FRAMEWORK OF EVALUATION OF ACTIVITY CENTRE IMPLEMENTATION (AUTHOR) ...... 70

FIGURE 9 - OVERVIEW OF THE METHOD USED TO CREATE A LAND USE DATABASE ...... 79

FIGURE 10- EXAMPLE GSV IMAGE (GOOGLE STREET VIEW) ...... 81

FIGURE 11 - RELATIVE CHANGE IN NET POPULATION AND DWELLING DENSITIES, 1996 TO 2016 ...... 127

FIGURE 12 - CENTRE LOCATIONS WITH DENSITY INTENSIFICATION SCORES ...... 133

FIGURE 13 - PROPORTIONS OF DWELLING TYPES IN AND OUTSIDE CENTRES, 1996 & 2016, BY LOCATION ...... 135

FIGURE 14 - CENTRE LOCATIONS WITH DWELLING MIX INTENSIFICATION SCORES ...... 141

FIGURE 15 - NET EMPLOYMENT DENSITY IN 1996 AND 2016 ...... 144

FIGURE 16 - PERCENTAGE ESTIMATED EMPLOYMENT CHANGE BY USE, TOP 5, 1996-2016 ...... 147

FIGURE 17 - EMPLOYMENT PLOT RATIO, 1996 AND 2016 ...... 149

FIGURE 18 - CENTRE LOCATIONS WITH EMPLOYMENT INTENSIFICATION SCORES ...... 151

FIGURE 19 - PROPORTIONS OF ESTIMATED LAND USE BY CENTRE, 2016 AND 1996 ...... 155

FIGURE 20 - EXAMPLE OF ACTIVE FRONTAGE "TRACING" IN LOGAN CENTRAL (IMAGE SOURCE: NEARMAP 2016) ...... 159

FIGURE 21 - ACTIVE FRONTAGE LENGTHS (METRES), 2016 AND 1996 ...... 160

FIGURE 22 - CENTRE LOCATIONS WITH CONFORMANCE CLASSIFICATION ...... 167

FIGURE 23 - CHERMSIDE 1997 (LEFT) AND 2016 (RIGHT) – (THE STATE OF QUEENSLAND, 2018), (NEARMAP 2016) ...... 169

FIGURE 24 - HIGH DENSITY DWELLINGS IN CHERMSIDE (GOOGLE STREET VIEW 2016)...... 170

FIGURE 25 - A NEW MIXED USE OFFICE BUILDING IN CHERMSIDE (GOOGLE STREET VIEW 2017) ...... 170

FIGURE 26 - CHERMSIDE - CHANGED USES AND POINTS OF INTEREST ...... 171

FIGURE 27 - CARINDALE 1997 (LEFT) AND 2016 (RIGHT) – SOURCE (THE STATE OF QUEENSLAND, 2018), (NEARMAP 2016) ... 172

FIGURE 28 - APARTMENTS NEXT TO WESTFIELD CARINDALE (LEFT), AND AN EXAMPLE OF THE MORE TYPICAL LOW DENSITY USES

(RIGHT) (GOOGLE STREET VIEW 2016) ...... 173

FIGURE 29 - CARINDALE - CHANGED USES AND POINTS OF INTEREST ...... 174

FIGURE 30 - IPSWICH 1997 (LEFT) AND 2016 (RIGHT) – SOURCE (THE STATE OF QUEENSLAND, 2018), (NEARMAP 2016) ...... 175

FIGURE 31 - IPSWICH'S NEW OFFICE TOWER AND TRADITIONAL MAIN STREET ...... 176

FIGURE 32 - HOUSE TO OFFICE CONVERSIONS (LEFT) AND THE OAKS ASPIRE RESIDENTIAL TOWER (RIGHT) (GOOGLE STREET VIEW 2017) ...... 176

FIGURE 33 - IPSWICH - CHANGED USES AND POINTS OF INTEREST ...... 178

FIGURE 34 - SPRINGWOOD 1996 (LEFT) AND 2016 (RIGHT) – SOURCE (THE STATE OF QUEENSLAND, 2018), (NEARMAP 2016)179

FIGURE 35 - SPRINGWOOD TOWERS HOTEL BEHIND A NEW MIXED-USE OFFICE (LEFT), AND A NEW SMALL SCALE OFFICE ON

CINDERELLA DRIVE (RIGHT) (GOOGLE STREET VIEW 2018) ...... 180

FIGURE 36 - SPRINGWOOD - CHANGED USES AND POINTS OF INTEREST ...... 181

FIGURE 37 - PLANNING REGULATIONS AND THEIR TYPE OF REFERENCE TO REGIONAL POLICY, 1996, 2006, AND 2016 ...... 186

FIGURE 38 - DIFFERENCES IN LAND AREA OF RESIDENTIAL DIS BETWEEN 1996 AND 2016, AS A PERCENTAGE OF TOTAL LAND AREA, BY

LOCATION ...... 191

FIGURE 39 - DIFFERENCES IN LAND AREA OF RESIDENTIAL DIS BETWEEN 1996 AND 2016, AS A PERCENTAGE OF TOTAL LAND AREA, BY

CENTRE ...... 192

FIGURE 40 - DIFFERENCES IN LAND AREA OF COMMERCIAL DIS BETWEEN 1996 AND 2016, AS A PERCENTAGE OF TOTAL LAND AREA,

BY LOCATION ...... 194

FIGURE 41 - DIFFERENCES IN LAND AREA OF COMMERCIAL DIS BETWEEN 1996 AND 2016, AS A PERCENTAGE OF TOTAL LAND AREA,

BY CENTRE ...... 195

FIGURE 42 - DIFFERENCES IN LAND AREA OF BULKY GOODS RETAIL DIS BETWEEN 1996 AND 2016, AS A PERCENTAGE OF TOTAL LAND

AREA, BY LOCATION ...... 196

FIGURE 43 - DIFFERENCES IN LAND AREA OF BULKY GOODS RETAIL DIS BETWEEN 1996 AND 2016, AS A PERCENTAGE OF TOTAL LAND

AREA, BY CENTRE ...... 197

FIGURE 44 - DIFFERENCES IN LAND AREA OF INDUSTRIAL DIS BETWEEN 1996 AND 2016, AS A PERCENTAGE OF TOTAL LAND AREA, BY

LOCATION ...... 198

FIGURE 45 - DIFFERENCES IN LAND AREA OF INDUSTRIAL DIS BETWEEN 1996 AND 2016, AS A PERCENTAGE OF TOTAL LAND AREA, BY

CENTRE ...... 199

FIGURE 46 - PERCENTAGE OF LAND PERMISSIVE OF MIXED USE DEVELOPMENT IN 1996 AND 2016, BY LOCATION ...... 200

FIGURE 47 - PERCENTAGE OF LAND PERMISSIVE OF MIXED USE DEVELOPMENT IN 1996 AND 2016, BY CENTRE ...... 201

FIGURE 48 - BROWNS PLAINS LAND USE CONFORMANCE, 1996 AND 2016 ...... 206

FIGURE 49 - SPRINGWOOD LAND USE CONFORMANCE, 1996 AND 2016 ...... 207

FIGURE 50 - DEVELOPMENT CONFORMANCE WITH LAND USE REGULATIONS BY LOCATION, 1996 TO 2016 ...... 209

FIGURE 51 - DEVELOPMENT CONFORMANCE WITH LAND USE REGULATIONS BY CENTRE, 1996 TO 2016 ...... 210

FIGURE 52 - TYPE OF USE CHANGE BY LOCATION, 1996 TO 2016 ...... 211

FIGURE 53 – PERCENTAGE OF EXCEEDING SITES, BY TYPE OF USE, IN OUTER LOCATIONS, 1996 TO 2016 ...... 213

FIGURE 54 - ARTIST'S IMPRESSION OF THE FUTURE OF SPRINGWOOD (LOGAN CITY COUNCIL, 2009) ...... 238

FIGURE 55 - PROPORTIONS OF DWELLING TYPES, 2016 AND 1996 ...... 287

List of Tables

TABLE 1 – SUMMARY OF SUSTAINABILITY BENEFITS OF COMPACT URBAN FORMS (OECD, 2012, P. 57) ...... 26

TABLE 2 – MATRIX OF CONFORMANCE VS. PERFORMANCE IN EVALUATIVE APPROACHES (AUTHOR) ...... 53

TABLE 3 - DESCRIPTION OF REGIONAL ACTIVITY CENTRE TYPES (THE STATE OF QUEENSLAND, 2016B) ...... 74

TABLE 4 - SELECTED ACTIVITY CENTRES FOR EVALUATION ...... 75

TABLE 5 - RESULTS OF RELIABILITY TESTING OF GSV LAND USE DATABASE OBSERVATIONS ...... 84

TABLE 6 - DWELLING TYPE CLASSIFICATIONS ...... 88

TABLE 7 - CENSUS EMPLOYMENT VS. FOOTPRINT ESTIMATED EMPLOYMENT ...... 94

TABLE 8 - BUILDING FRONTAGE CLASSIFICATIONS ...... 99

TABLE 9 - SUMMARY OF COMPACT ACTIVITY CENTRE INDICATORS ...... 104

TABLE 10 - DEVELOPMENT INTENSITY SCORE (DIS) CRITERIA ...... 110

TABLE 11 - RESULTS OF RELIABILITY TESTING OF DIS CODING ...... 112

TABLE 12 - EXISTING COMPACTNESS VARIABLES ...... 118

TABLE 13 - PROPERTY VARIABLES...... 119

TABLE 14 - TRANSPORT VARIABLES ...... 120

TABLE 15 - PLANNING POLICY VARIABLES ...... 121

TABLE 16 - SOCIO-ECONOMIC VARIABLES ...... 122

TABLE 17 - INDICATORS TO MEASURE RESIDENTIAL DENSITY OBJECTIVES ...... 126

TABLE 18 - DIFFERENCE IN PERCENTAGE POINTS BETWEEN RELATIVE BASELINE AND CENTRE CHANGES TO POPULATION, DWELLINGS

AND DENSITIES, 1996 TO 2016 ...... 128

TABLE 19 - CHANGE IN AVERAGE LAND AREA PER LOW DENSITY DWELLING, 1996-2016 (ORDER BY RELATIVE CHANGE) ...... 129

TABLE 20 - CHANGE IN THE PERCENTAGE OF POPULATION LIVING IN LOW DENSITY DWELLINGS, 1996-2016, INCLUDING BASELINE

AREAS ...... 131

TABLE 21 - AVERAGE DENSITY SCORE IN 1996, 2016, AND AVERAGE DENSITY INTENSIFICATION SCORE 1996 TO 2016 ...... 132

TABLE 22 - INDICATORS OF DWELLING MIX ...... 134

TABLE 23 - RELATIVE CHANGE IN DWELLING TYPES, 1996-2016 ...... 137

TABLE 24 - INDEX OF QUALITATIVE VARIATION OF DWELLING TYPES, 2016 AND 1996 ...... 138

TABLE 25 - DWELLING MIX INDICATORS FOR 2016 AND 1996, AND AVERAGE DWELLING MIX INTENSIFICATION SCORE ...... 140

TABLE 26 - INDICATORS OF EMPLOYMENT ...... 142

TABLE 27 - ESTIMATED EMPLOYMENT AND RELATIVE CHANGE, 1996 TO 2016 ...... 145

TABLE 28 - AVERAGE EMPLOYMENT INTENSITY BY CENTRE, 1996 AND 2016 ...... 148

TABLE 29 - OVERALL EMPLOYMENT SCORES FOR 1996 AND 2016, AND EMPLOYMENT INTENSIFICATION SCORE ...... 150

TABLE 30 - INDICATORS OF MIXED USE AND ACCESSIBILITY ...... 152

TABLE 31 - LAND USE VARIATION, 1996 AND 2016 ...... 154

TABLE 32 - DIFFERENCE IN PROPORTIONS OF AREA OF ESTIMATED LAND USE FROM 1996 TO 2016...... 156

TABLE 33 - AVERAGE EUCLIDEAN DISTANCE, 2016 AND 1996 ...... 157

vii

TABLE 34 - INDICATORS OF RESIDENTIAL PROXIMITY, 2016 AND 1996 ...... 158

TABLE 35 - ACTIVE FRONTAGES AS PROPORTION OF ALL COMMERCIAL FRONTAGES, 2016 AND 1996 ...... 161

TABLE 36 - OVERALL MIXED USE SCORES FOR 1996 AND 2016 ...... 162

TABLE 37 - OVERALL COMPACTNESS SCORES, 1996 AND 2016, AND OVERALL INTENSIFICATION SCORE ...... 163

TABLE 38 - OVERALL CENTRE INTENSIFICATION SCORES, 1996 TO 2016 ...... 164

TABLE 39 - CLASSIFICATION OF CENTRE CONFORMANCE ...... 165

TABLE 40 - SUMMARY OF CHANGES TO LAND USE REGULATIONS BY CENTRE, 1996 TO 2016 ...... 190

TABLE 41 - CLASSIFICATION OF LAND USE CONFORMANCE TO LAND USE REGULATIONS AS A PERCENTAGE OF CENTRE AREA, 1996 & 2016 ...... 204

TABLE 42 - RELATIONSHIP BETWEEN CENTRE INTENSIFICATION AND EXISTING COMPACTNESS IN 1996 ...... 217

TABLE 43 - RELATIONSHIP BETWEEN CENTRE INTENSIFICATION AND UNIT PRICES ...... 218

TABLE 44 -RELATIONSHIP BETWEEN CENTRE INTENSIFICATION AND TRANSPORT FACTORS ...... 219

TABLE 45 - RELATIONSHIP BETWEEN CENTRE INTENSIFICATION AND PLANNING SCHEME CHANGE FACTORS ...... 220

TABLE 46 - RELATIONSHIP BETWEEN CENTRE INTENSIFICATION AND OTHER PLANNING POLICY VARIABLES ...... 220

TABLE 47 - RELATIONSHIP BETWEEN CENTRE INTENSIFICATION AND THE 1996 INDEX OF EDUCATION AND OCCUPATION (ABS) .... 221

TABLE 48 - RELATIONSHIP BETWEEN EXISTING COMPACTNESS INDICATORS AND OTHER VARIABLES ...... 223

TABLE 49 - RELATIONSHIP BETWEEN PROPERTY, SOCIO-ECONOMIC, AND TRANSPORT VARIABLES ...... 224

TABLE 50 - RELATIONSHIPS BETWEEN CENTRE INTENSIFICATION AND PROPERTY, SOCIO-ECONOMIC AND TRANSPORT VARIABLES WHILE

CONTROLLING FOR EACH OF THESE VARIABLES ...... 225

TABLE 51 - JOURNEY TO WORK MODE SHARE IN CHERMSIDE, 1996 AND 2016 ...... 246

TABLE 52 - RATIONALE FOR LOCATING CENTRAL TRANSIT NODES IN EACH CENTRE ...... 279

TABLE 53 - 1996 AND 2016 POPULATION AND DWELLING DENSITIES EXPRESSED IN NET RESIDENTIAL HECTARES ...... 283

TABLE 54 - ESTIMATED CENTRE POPULATION AND DWELLING NUMBERS IN 2016 AND 1996 ...... 283

TABLE 55 - RELATIVE CHANGE TO BASELINE POPULATION, DWELLINGS, AND DENSITIES 1996 TO 2016 ...... 284

TABLE 56 - RELATIVE CHANGE IN POPULATION AND DWELLING DENSITIES 1996 -2016 IN TERMS OF BUILT-UP HECTARES ...... 284

TABLE 57 - RESIDENTIAL DENSITY INDICATORS FOR 2016 AND 1996 EXPRESSED AS Z-SCORES (ORDERED BY 2016 SCORE) ...... 285

TABLE 58 - DIFFERENCE IN RESIDENTIAL DENSITY INDICATORS, AND RELATIVE ESTIMATED POPULATION AND DWELLING CHANGE 1996

TO 2016, EXPRESSED AS Z-SCORES (ORDERED BY SCORE) ...... 286

TABLE 59 - CHANGE IN PROPORTIONS OF DWELLING TYPES, 1996 TO 2016 ...... 288

TABLE 60 - ABSOLUTE CHANGE IN DWELLING NUMBERS BY TYPE OF DWELLING, 1996 TO 2016 ...... 288

TABLE 61 - DWELLING MIX INDICATORS FOR 2016 AND 1996 EXPRESSED AS Z-SCORES (ORDERED BY 2016 SCORE)...... 289

TABLE 62 - RELATIVE CHANGE OF DWELLING TYPES (1996-2016) EXPRESSED AS Z-SCORES (ORDERED BY OVERALL AVERAGE Z-SCORE) ...... 290

TABLE 63 - NET EMPLOYMENT DENSITY, 1996 AND 2016 ...... 290

TABLE 64 - MOST SIGNIFICANT ESTIMATED EMPLOYMENT CHANGE BY USE AND CENTRE, 1996 TO 2016 ...... 291

TABLE 65 - EMPLOYMENT PLOT RATIO, 2016 AND 1996, (ORDERED BY DIFFERENCE) ...... 292

TABLE 66 – EMPLOYMENT INDICATORS FOR 2016 AND 1996 EXPRESSED AS Z-SCORES (ORDERED BY 2016 Z SCORE) ...... 293

viii

TABLE 67 - DIFFERENCE IN EMPLOYMENT INDICATORS, AND RELATIVE ESTIMATED EMPLOYMENT CHANGE 1996 TO 2016, EXPRESSED

AS Z-SCORES (ORDERED BY SCORE) ...... 293

TABLE 68 - MIXED USE INDICATORS 2016 AND 1996, EXPRESSED AS Z SCORES AND ORDERED BY 2016 SCORE ...... 294

TABLE 69 - DIFFERENCE IN MIXED USE INDICATORS EXPRESSED AS Z-SCORES (ORDERED BY SCORE) ...... 295

TABLE 70 – OVERALL CENTRE COMPACTNESS SCORES, 2016 AND 1996 ...... 295

TABLE 71 - SUMMED LAND AREA (IN HECTARES), BY CENTRE, OF 2016 RESIDENTIAL DIS ...... 316

TABLE 72 - SUMMED LAND AREA (IN HECTARES), BY CENTRE, OF 2006 RESIDENTIAL DIS ...... 316

TABLE 73 - SUMMED LAND AREA (IN HECTARES), BY CENTRE, OF 1996 RESIDENTIAL DIS ...... 317

TABLE 74 - SUMMED LAND AREA (IN HECTARES), BY CENTRE, OF 2016 COMMERCIAL DIS ...... 317

TABLE 75 - SUMMED LAND AREA (IN HECTARES), BY CENTRE, OF 2016 INDUSTRIAL DIS ...... 319

TABLE 76 - SUMMED LAND AREA (IN HECTARES), BY CENTRE, OF 2006 INDUSTRIAL DIS ...... 319

TABLE 77 - SUMMED LAND AREA (IN HECTARES), BY CENTRE, OF 1996 INDUSTRIAL DIS ...... 320

TABLE 78 - SUMMED LAND AREA (IN HECTARES), BY CENTRE, OF 2016 BULKY GOODS RETAIL DIS ...... 320

TABLE 79 - SUMMED LAND AREA (IN HECTARES), BY CENTRE, OF 2006 BULKY GOODS RETAIL DIS ...... 321

TABLE 80 - SUMMED LAND AREA (IN HECTARES), BY CENTRE, OF 1996 BULKY GOODS RETAIL DIS ...... 321

TABLE 81 - CLASSIFICATION CRITERIA OF CONFORMING, UNDER-DEVELOPED, OR EXCEEDING USES ...... 322

TABLE 82 - CORRELATION MATRIX OF INDEPENDENT VARIABLES ...... 324

Statement of Original Authorship

The work contained in this thesis has not been previously submitted to meet requirements for an award at this or any other higher education institution. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made.

QUT Verified Signature

xii

Acknowledgements Although the proceeding statement of authorship acknowledges that this thesis is my original work, it is also a work that would not have been possible were it not for the kind support of so many other people. To begin, I will honour a promise I made in the thick of hunting historical documents and pronounce my love for the librarians of the world; all of them. I firstly dedicate this thesis to you. This research required the assistance of librarians and archivists from all manner of institution. To a person, they consistently applied indomitable professionalism as they diligently sought information on my behalf. Whether relentlessly pursuing the faint scent of some obscure, coffee stained document or physically hauling the hundredth planning scheme map from a long-ignored repository, they never once questioned my purpose or balked at the required effort. This thesis is dependent on the unflinching application of their sacred duty to information. These attributes are perfectly exemplified in QUT’s master information hunter, Sharron Stapleton, who was there for me whenever called upon.

To my supervisors, it is no exaggeration to say that your guidance and support is what made this research possible. The husband and wife super team of Severine Mayere and Paul Donehue gently and compassionately held my course to the finish and coaxed me through doldrums of self-doubt and uncertainty. Their support extended far beyond the research itself and enabled me to proceed without impoverishing my family. I am also indebted to a true scholar, Carl Grodach, for taking me in when this research was just a crude proposal, and for encouraging me to think far beyond the confines of my conceptualisations formed from professional practice.

To my friends, family, and colleagues too numerous to name. To my bandmates in the Toxic Bears. You were there with me throughout. Your kind words of encouragement, your advice, and empathetic listening provided me with more support than you might think. I needed you. Your ongoing consideration was more than just comforting, it subtly embedded itself into the research and my writing, both of which are now the better for it.

I feel deeply honoured and privileged to have had the opportunity to pursue a PhD. The experience was all encompassing and was as humbling as it was enlightening. I count myself as being extremely fortunate to have had you all with me, and I sincerely thank every one of you. I must also acknowledge the generous support from QUT in providing me with a CEBE scholarship for the duration of my research; without which, this research would never have proceeded.

Most significantly however, I acknowledge the extraordinary demands imposed on my beloved Michelle, and my young sons Matthew and Andy. For you, my research resulted in everything from mountainous piles of laundry, to missed bedtime stories and backyard rugby matches. Despite this you never complained, accepting that this is what I needed to do. I cherished every one of those sleepy welcome home hugs when I’d disturb your slumber for a goodnight kiss. The months where I was absent from your lives, the late nights, and the many weekends I spent with this thesis instead of you has resulted in my most significant research finding; the reaffirmation that I treasure you more than anything else in the world. I thrive from the pursuit of knowledge, but you will now always come first.

xiii

xiv

Abstract

Australian planning authorities have been attempting to develop more compact cities for several decades with the intention of realising a range of urban sustainability benefits. In most capital cities, these attempts take the form of metropolitan level plans which seek to create an intense urban core, supported by a network of compact activity centres throughout the middle and outer rings of the broader conurbation. Existing research suggests that these activity centre policies are failing to be implemented, particularly in car-dependent outer areas where further compactness could potentially yield the greatest benefits. However, these existing studies have been drawn from few cases, often use data that is now more than ten years old, and do not make use of theories of plan evaluation.

This research builds on existing studies to evaluate activity centre planning in terms of both plan outcomes and use, and explores the relationships between plan implementation and a range of explanatory factors. Undertaking plan evaluation of this nature is critical to inform future planning efforts to improve urban sustainability. This research draws on theories of plan evaluation to comprehensively evaluate the conformance and performance of greater Brisbane’s compact activity centre policies and considers the factors that influence their development. In order to overcome a number of data limitations, the research develops a detailed land use database derived from observations of Google Street View and aerial images of more than 26,000 individual sites, which is then used to interpolate secondary data sources and comprehensively evaluate plan conformance over a twenty year period. The research then evaluates how local governments made use of activity centre policy when developing and applying land use regulations through the analysis and quantification of the regulatory intent of current and historical planning documents. Finally, the research uses statistical correlations between measures of centre intensification and a range of potential influencing factors such as changes to the permissiveness of land use regulations, public transport accessibility, distance to the CBD, property values, and socio-economic status.

The results reveal that despite strong performance in the use of activity centre policy by local governments, policy conformance with its intended outcomes was poor. The permissiveness of land use regulations had little to no impact on centre implementation with intensification being most strongly related to socio-economic, property price, and employment factors. The results indicate that policy makers need to reconsider the types of centres that are suitable for future compaction, as well as their reliance on implementation mechanisms based on changing the regulatory land use planning system in the hope of attracting private sector development. The research raises doubts about the Australian planning system’s capability to deliver on its core promises and reshape urban development. These results have implications for whether compact activity centre policy can be a viable mechanism to achieve its purported sustainability benefits.

1. Introduction

This research evaluates the implementation of 20 years of compact activity centre policy in the greater Brisbane area. In-line with plan evaluation theory, the policy is evaluated in terms of its outcomes as well as its use. This chapter provides a summary of background information to identify how the research aligns with current scholarship on this topic. From this, the research questions are then described, before providing an overview of the approach and methods used to address them. The benefits of the research are also summarised, and the chapter concludes with an outline of the thesis’s contents.

1.1. Background

Metropolitan planning policy in Australia has included a focus on creating more compact cities for the past twenty years. These policies typically propose fundamental changes to the existing urban form to create a polycentric network of compact activity centres (Dodson, 2012; Forster, 2006). In line with standard rationales for the compact city, policy makers argue that doing so will result in a range of environmental, economic, and social benefits (Freestone, 2012). These purported sustainability benefits are far reaching and promise to improve all manner of urban issues including climate change prevention through reductions in CO2 emissions from vehicle trips, the conservation of undeveloped areas due to a reduced need for sprawling forms of development, increases in productivity from clustering of businesses and shorter commuting times for workers, the more efficient use of existing infrastructure, a range of equity benefits such as more affordable housing, improved access to employment and services, and improved health outcomes from increases in active transportation use (OECD, 2012).

Activity centre policies have proven to be a popular and politically expedient way to compartmentalise the imperative for more compact cities by minimising the threat of change to more established suburban neighbourhoods. As a result, compact activity centre policy is now ubiquitous across Australian cities. Most attention has been focussed on the long standing centre policies of Sydney and Melbourne, however Adelaide, Perth and Canberra have also had multi-nodal centre policies in place since the 1960s, with Brisbane developing a hierarchical centre policy in the 1980s (Quirk, 2007; The State of Queensland, 1993b). In the 1990s, these commercial centre policies began to take the form of today’s planned activity centres through the incorporation of compact city principles. This change was typically characterised by plans to integrate higher density residential uses with commercial centre

16 policies in order to improve sustainability (Dodson, 2012; Forster, 2006). This trajectory is clearly apparent in activity centre policy for greater Brisbane which increasingly linked sustainability outcomes to centre development across its various iterations of regional level planning (The State of Queensland, 1995, 2005c, 2009, 2017c). The plans reasoned that centres characterised by higher residential densities, a greater diversity of housing types, and more mixed clusters of employment generating uses, would lead to the types of sustainability benefits supposedly inherent in more compact urban forms. These general principles for compact activity centres are now common in metropolitan planning policy in most Australian cities (Forster, 2006).

However, a number of authors, both in Australia and internationally, raised concerns about whether compactness policies would in fact result in the purported sustainability outcomes, and predicted that the policies to reshape existing urban forms would prove difficult to implement (Birrell et al., 2005; Breheny, 1997; Gordon & Richardson, 1997; Troy, 1996; Williams, 1999). More recent empirical studies in Australia tend to support these early predictions and suggest that limited progress has been made towards their implementation, especially in middle and outer suburban contexts. Just as the initial commercial centre policies failed to be influential in urban development (McLoughlin, 1992), evidence suggests that the updated activity centre policies are also having little effect on the evolution of urban forms, either in terms of housing development (Chhetri et al., 2013; Newton & Glackin, 2014; Phan et al., 2009), or employment clustering (Day et al., 2015).

There are a range of limitations associated with existing studies which prompt the need for further research. Existing empirical studies are almost entirely Melbourne-centric, and expanding the research to include additional cases would begin to address whether issues associated with the implementation of compact activity centre policies are the result their localised application or due to more fundamental issues inherent in planning for urban development in Australian cities. Much of the existing research evaluating the progress of activity centres is based on data that is now ten years old, thereby omitting the significant growth in infill residential development that has occurred more recently in Australian cities. Although activity centre plans typically involve planning horizons of 20 years or more, most policy evaluations to date have involved observations over shorter timeframes of less than five years (BITRE, 2013; Chhetri, et al., 2013; Phan, et al., 2009; The State of Queensland, 2006b, 2008). This raises concerns as to whether current evaluations have permitted sufficient time for plans to take effect. Longer term comparisons of discrete urban areas are inhibited by a range of data limitations, particularly in terms of key population and housing data from

17 censuses, prompting researchers to consider alternative methods and data sources on which to base their research (Buxton & Tieman, 2005; Coffee et al., 2016).

Of more conceptual concern, existing plan implementation evaluations of compact activity centres are based on planned outcomes, and do not consider plan use. Plan evaluation theorists consider such distinctions in approach to be important factors when framing evaluative research, with persistent calls for evaluation to make use of combined approaches (Alexander, 2006; Alexander & Faludi, 1989; Oliveira & Pinho, 2010a). The current “conformance” based approaches (i.e. evaluations of planned outcomes) fail to consider whether plans are “performing” (being used when making decisions). Without such considerations, it is not possible to determine whether conformance failures are occurring due to the plan not being used as intended, or the result of other factors. To date, there is limited empirical work that considers the reasons associated with activity centre implementation. An up to date, holistic evaluation of compact activity centre policy, that overcomes the extant data issues, and which uses a research framework informed by plan evaluation theory would address these deficiencies and add further knowledge to how activity centre policy is faring, and its prospects for achieving its desired sustainability goals.

1.2. Research questions

This research seeks to address three key questions in relation to the implementation of activity centre policy in greater Brisbane:

.1. How have greater Brisbane’s activity centres changed in-line with compact city based metropolitan policy?

.2. How has metropolitan planning policy for compact activity centres influenced land use regulations, and how do actual land use changes conform to these regulations?

.3. What are the relationships between commonly cited explanatory factors for activity centre conformance with the achievement of compact activity centre objectives?

The first question evaluates activity centre policy in terms of conformance. This approach involves comparing planned activity centre outcomes with actual changes to land use and the built form over a 20 year period. Addressing this question provides the first detailed, long term evaluation of the outcomes of activity centre policy in greater Brisbane, thereby adding depth to the relatively small number of studies that specifically examine the implementation of activity centre policy more broadly. The second question combines both conformance and

18 performance based approaches to plan evaluation. Regional level activity centre planning is primarily intended to be implemented by aligning the regulatory land use planning system with regional planning intent. Examining the changes to local government land use regulations over time, as well as how well actual land use changes align with these regulations, provides an indication of plan performance; i.e. the degree to which the activity centre policy has been used by key actors involved in its implementation. The relationship between regional level activity centre policy and its application to land use regulations has yet to be considered by existing research, and enables an investigation of whether the results from the first research question are related to how plans are used, or due to other factors. This analysis is the key purpose of the third research question which seeks to explore the relationship between policy implementation, policy use, and a range of other factors. A range of existing papers propose factors that influence the implementation of compact activity centre policy. However, there has been little empirical research to date that examines the connections between these factors and the results of holistic evaluations of plan implementation.

Compact activity centre policy is a key component of land use strategies to improve urban sustainability. Addressing these research questions is necessary to inform future planning strategies and research.

1.3. Research approach

This research is an evaluation that considers the effects of policy implementation in terms of land use and demographic change, and in terms of changes to subordinate policy documents and regulations. As such, it primarily makes use of quantitative research methods of data collection and analysis. The overall research approach to evaluation is mostly positivist in nature. The research considers plans and policies to be effectuating devices and that, after a sufficient period (in this instance two decades), it is reasonable to expect evidence of the plans’ intended effects.

This is particularly the case in the approach used to address the first research question. This question examines whether compact activity centre policy is conformant; i.e. did the policy result in the desired outcomes. To address this question it is necessary to identify the intended outcomes of activity centre policy, and then seek evidence of these outcomes in actual changes to the nominated centres. A review of greater Brisbane’s activity centre policies revealed that the policy intent has been remarkably consistent over 20 years, in both spatial and substantive terms. According to activity centre policy, the broader Brisbane conurbation is to develop into multi-nodal network of activity centres characterised by higher residential densities, a greater

19 diversity of dwelling types, and include greater opportunities for employment and service provision through an increased mixes of use types. These objectives are similar to those typically associated with more general policies to create compact cities. A review of literature that measured city compactness identified a range of potential indicators that could be used to measure compact activity centre change. The nature of activity centre policy requires indicators to be able to consistently measure change to discrete urban areas over the 20 year study period. After a consideration of data availability and the suitability of these indicators to measure activity centre policy, a collection of 15 separate indicators were selected. There are a number of complications that severely limit the suitability of existing data sources for this purpose, which prompted the development of a new method of data collection to enable the longitudinal measurement of centre change (Limb et al., 2018). Using observations from Google Street View and aerial imagery, a detailed land use database was developed to track land use change on every site in the study area over the 20 year period. When combined with a range of secondary data sources, the land use database allows for the comparison of centres using the selected compactness indicators at the beginning of the activity centre policy and after a period of 20 years. The change in compactness is then compared between each centre, as well as between a selection of “baseline” measures that compare centre change to change in the broader conurbation. The degree of plan conformance is then assessed on the basis of these comparisons.

Some plan evaluation theorists believe such positivist forms of evaluation reflect an outdated view of planning which conceives of plans as “blueprints” rather than decision making instruments (Alexander & Faludi, 1989; Altes, 2006; Faludi, 2006). These scholars argue that plan success should be evaluated based on whether the plan was used as intended when making key decisions, rather than in terms of whether it achieved its planned outcomes, even if those decisions result in the plan not being followed. However, others note that regardless of whether planning theorists no longer consider plans as blueprints, those involved in the development of plans very much intend them to be effectuating devices, and failing to consider whether a plan achieves its intended outcomes therefore allows policy and decision makers to avoid taking responsibility (Loh, 2011; Talen, 1997). The approach used in this research is sympathetic to this latter view, however it also accepts that understanding how a plan was used is an important consideration in order to form a more comprehensive evaluation of policy. This is reflected in the second research question which seeks to determine whether activity centre policy was used when making decisions to update regulatory plans, and to approve development. Knowing if activity centre policy was used as intended helps to understand if the results of the first question were due to their performance (i.e. the extent to which they were used) or due to other factors.

Plan performance evaluations are typically undertaken either by interviewing and/or surveying key policy actors, or through the analysis of relevant documents. The key implementation mechanism for greater Brisbane’s compact activity centre policy involves changing local government land use regulations to reflect the intended centre outcomes. Evaluating plan performance in this instance therefore requires access to the actors and/or documents associated with these regulatory changes. Due to the 20 year time period and the large number of local governments involved, several of which no longer exist, it was considered impractical to identify and consult those involved in making and amending planning regulations. The regulations themselves can be compared over time to see if there is evidence of activity centre policy being used. Finding specific reference to activity policy is one indication of its use. However, the material components of land use regulations are the regulations themselves. It is therefore necessary to examine a complex mix of different parts of the regulatory documents in order to determine how they impact individual development sites. The research therefore codes these parts using a development intensity score (DIS) which can then be applied to quantify the permissiveness of land use regulations on any land in the study area. The DIS are aligned with key use types associated with compact activity centre intentions being, residential, commercial, industrial, and bulky goods retail uses. By comparing changes to DIS over time, it is possible to determine the degree to which regulations have been changing either with, or contrary to, the intend policy. The DIS also permit the comparison of regulatory intent to actual land use change in order to measure whether decisions to approve land use changes are occurring within the bounds anticipated by the regulations in force at the time of the land use change. The results of this analysis allow for the extent of plan performance in each centre, to be deduced.

The final research question seeks to explore possible explanations for the results observed from the first. This is undertaken using statistical correlations to determine the strength of relationship between the degree of centre intensification (determined in the first question), and measures of a range of other factors. These factors include property prices and price changes, public transport accessibility, socio-economic status, the degree of existing compactness, and changes to the permissiveness of planning policy. Suitable indicators for each of these factors are selected after a consideration of the limitations posed by the availability of historical data at scales appropriate for the analysis of individual centres. The most appropriate indicators often use ordinal scales, and the data across centres rarely has normal distributions. Non- parametric, spearman rank order correlations are therefore used to examine the relationship between the factors and centre intensification. Due to the relatively small number of nominated activity centres in the policy for greater Brisbane, it is not possible to reliably undertake more complex statistical analysis that determines causation. Partial correlations are

21 therefore also performed to control for the influence of the independent variables on each other and provide further depth to the research findings.

1.4. Research outcomes

This research has implications for practice, theory and methods. State and local governments spend significant amounts of time and resources attempting to implement compact city policies through the land use planning system, and other mechanisms. Evaluating the implementation and use of these plans will advance knowledge of conditions under which such policies are most likely to be implemented. Such knowledge is likely to be beneficial to future policy makers developing similar land use plans, and to the achievement of their purported sustainability benefits. Plan evaluation theorists continue to call for further research evaluating the implementation of plans. This research will contribute to understandings of plan evaluation by providing further examples that consider the relationships between plan conformance and plan performance. The research will also enable further contributions to the ongoing debates about the purported sustainability benefits of more compact urban forms through the development of appropriate methods to measure changes to discrete areas and their relative compactness, as well as methods of measuring plan use and performance. Such understandings are a necessary precondition to identify plan success and failure in order to subsequently determine whether more compact urban forms do in fact provide more sustainable outcomes, and thus guide how policy makers can best improve urban sustainability issues.

1.5. Dissertation Outline

This dissertation provides a detailed description of the research problem and its context, reviews relevant literature, and describes a research design for how the research was undertaken before presenting the research results and a discussion of their implications.

Chapter 2 provides relevant historical context and background to the topic. It outlines the rise and evolution of compact activity centre policy in Australia, reviews common rationales for the compact city, and the ongoing debates surrounding its efficacy and feasibility in achieving its supposed sustainability benefits. It also provides an in-depth review of the development of compact activity centre policy in the selected study area of greater Brisbane. This background is used to develop an appropriate evaluation framework and indicators of conformance and performance.

Chapter 3 reviews key literature on plan evaluation, examines the existing evaluations of compact activity centre policy, and considers the range of methods currently used in studies of plan implementation evaluation. Theories of plan evaluation provide a useful framework from which to examine activity centre policy in terms of plan use, and outcomes. Existing studies suggest that compact activity centre policy has often been poorly implemented, however an analysis that addresses a range of limitations is required to confirm this conclusion. The chapter also identifies the key methodological issues associated with plan evaluation research. The findings from this chapter are used to inform the development of a research plan.

Chapter 4 provides a research plan that includes the research questions and objectives, and a detailed methodology for how the research will be undertaken. Each research question was approached using a different series of methods. A new method was developed to address the first question that develops a historical land use database from Google Street View and aerial image observations. The database is used directly, as well as in the areal interpolation of secondary data sources to provide a suitable dataset from which long term observations and measurements of residential density, dwelling mix, employment, and mixed use, can be drawn. The second question is addressed through content analysis and a quantification of land use regulations to assess the degree to which local governments used activity centre policy when developing and applying their land use plans. The final question is addressed using statistical correlations to determine the strength of the relationships between the different variables.

Chapters 5, 6 and 7 present and discuss the findings for each research question respectively. The research finds that policy implementation has conformed poorly to its intended outcomes even though the policy performed strongly in influencing local government development and application of planning regulations. The poor conformance was primarily due to a lack of development activity in a large proportion of centres rather than from issues associated the nature or use of land use regulations. Regulatory change was subsequently shown to bare little relationship to centre intensification, which instead was most strongly associated with factors such as socio-economic status, property prices, and the nature of employment in each centre.

Chapter 8 discusses these results and concludes the dissertation. This discussion highlights the inherent implementation problem associated with existing activity centre policy. With property economics and existing employment being the driving force of centre intensification, the plan failed to account for these aspects when nominating the designated centres. Future plans therefore need to better consider implementation in terms of the nature of existing centres and the mechanisms through which change will occur. This result has implications for

23 the achievement of sustainability outcomes, and a range of future research is outlined to determine whether land use planning of this nature can provide viable sustainability benefits in the face of more immediate challenges such as preventing the risks of climate change.

2. The compact city, activity centres, and the evolution of planning for greater Brisbane

This research seeks to explore the implementation of compact activity centre policy, particularly in terms of the intended development of more polycentric urban forms. This chapter provides background and context about the compact city and the evolution of planning policy in Australia, with particular focus on policies for the greater Brisbane area. This review provides the necessary context from which to develop an evaluation framework to guide subsequent stages of the research. Background context is firstly provided to outline the conventional understanding of how urban policy has evolved in Australia to almost universally include strategies that develop a network of compact activity centres in metropolitan areas. These policies are heavily influenced by the compact city movement; a concept thats efficacy continues to be debated. The key rationale for the compact city and the arguments for and against are presented. This review finds that there is evidence that some sustainability objectives can be achieved by more compact urban forms and that the compact city is now ubiquitous in urban policy across the globe. Studying the implementation of compact city policy therefore provides useful insights into urban governance and planning, as well as establishing the preconditions necessary to further study issues of efficacy. This chapter establishes that Australian metropolitan plans have featured policies to reshape existing urban forms into a polycentric network of compact activity centres for more than two decades. Metropolitan planning for greater Brisbane is no exception. Here, policy makers have increasingly used sustainability principles to justify an urban form made up of a network of activity centres characterised by increased residential densities and greater housing diversity in proximity to clusters of mixed, employment generating uses.

2.1. The rationale and common critiques of the compact city

Australia’s metropolitan policy has been heavily influenced by the compact city movement. Compact city policies are commonly considered to have a range of sustainability benefits and are now ubiquitous in most OECD nations (OECD, 2012). The benefits of compact city forms are commonly considered in terms of their contribution to triple bottom-line sustainability

25 principles of environmental, economic, and social benefits as summarised in Table 1. Whether these benefits can actually be attributed to compact urban forms continues to be debated (Ewing & Hamidi, 2015). Some of the key arguments for and against the compact city are outlined in this section, which concludes that overall, current research demonstrates a link

Table 1 – Summary of sustainability benefits of compact urban forms (OECD, 2012, p. 57) between some sustainability benefits and the compact urban form. This study does not intend to directly contribute to this ongoing debate, but instead focuses on the implementation of compact city policies. Understanding the implementation of compact city policies is a key precondition to further research of the efficacy of the compact city.

The sustainability benefits of urban consolidation became a key driver for the adoption of a remarkably similar set of policies across Australian cities that aimed to create regional and metropolitan land use patterns that concentrated housing, employment and services in a poly- nodal network of compact centres (Dodson, 2012, p. 26; Forster, 2006, p. 178). Forster (2006, pp. 179-180) describes the normative case for these government polices as one that intends to “limit suburban expansion” and create a “strong multi-nuclear” settlement pattern, characterised by “high density housing around centres and transport corridors” that lead to improvements in energy, water and transport efficiencies, as well as more equitable access to housing. The alluring nature of such narratives helped the concept of urban consolidation to become “conventional wisdom” among planners by the mid-1990s (Breheny, 1997). Early

26 critics such as Troy (1996) challenged whether changing the urban form is a viable or desirable fix for the sustainability issues facing Australian cities. Other authors saw proposals for compact urban forms as a poorly evidenced idealistic attack on the highly popular suburban way of life (Gordon & Richardson, 1997), or rife with contradictions that substantially complicate implementation (Williams, 1999).

Urban consolidation policy in Australia was initially considered in terms of its economic benefits in reducing the costs to deliver services and infrastructure to urban populations, compared to the costlier proposition of servicing dispersed suburban settlements (Forster, 2006, p. 178; Searle, 2007, p. 4). This reasoning for the compact city is generally well accepted, with research from the United States confirming that sprawling urban forms are more costly to service with infrastructure than compact urban forms (Carruthers & Ulfarsson, 2003). Cervero (2001) highlighted correlations between higher population density and higher levels of employment productivity and economic performance. Such correlations have been explained by economic factors such as the concentration of knowledge and markets (Morikawa, 2011), and on efficiencies gained through improved transportation times for workers (Prud'homme & Lee, 1999).

Transportation is also a key factor in establishing the environmental benefits of the compact urban form. The work of Newman and Kenworthy (1989) proved to be seminal in establishing an environmental justification for urban consolidation. Newman and Kenworthy (1989) examined 32 cities across North America, Europe, Australia and Asia and determined that rates of gasoline consumption were substantially higher in cities with low population densities, even when controlling for price, income, and vehicle efficiency. The study concluded that increasing urban population densities and automobile dominance would therefore have environmental sustainability benefits by reducing emissions generated by automobiles (Newman & Kenworthy, 1989). The combined measures of traffic calming, public transit development, and focusing growth into densely populated “European style” “urban villages” were subsequently proposed as a solution to the negative effects of suburban sprawl (Newman & Kenworthy, 1992, 1999). Critics however did not believe such changes were the optimal method to achieve environmental benefits. Breheny (1995) accepted that more compact urban forms would reduce transport energy consumption, however the reduction would be relatively small and could be accomplished through means other than the significant, and potentially unviable, challenge of fundamentally altering well-established urban forms. Ewing and Hamidi (2015) identify a wide range of studies that show relationships between reduced vehicle miles travelled and compact cities. However, these studies often do not address increased travel time associated with congestion in more compact

27 areas. Kahn (2007) demonstrated that sprawling areas exhibited reduced travel times compared to compact areas, undermining some of the economic and environmental reasoning for the compact city. This suggests decentralised employment correlates with improved travel time more than population density, and Ewing and Hamidi (2015) concede that reductions in travel time in more decentralised areas counter the environmental improvements of reductions in distances travelled. Some research continues to find links between higher densities and reduced vehicle usage (Tanaguchi et al., 2008). However, other studies show that accessibility measures of active and public transport show greater correlations to mode change than population or employment density (Ewing & Cervero, 2010).

The other most common environmental benefits attributed to compact urban forms relate to reductions in energy use (and associated greenhouse gas emissions) and the preservation of natural areas and agricultural land. Andrews (2008) found that low-density, suburban forms resulted in higher emissions per capita than more dense urban locations. Other studies have found energy usage increases significantly in larger homes, and the prevalence of such homes is correlated with suburban city forms (Ewing & Rong, 2008). In Australia, the effects of urban form on pollution and energy efficiency were investigated by Newton (1997) whose models gave support for encouraging more dense urban forms such as compact urban core and multi- centred urban village or edge city concepts. Dodson (2012, pp. 26-27) asserts that evidence for the purported benefits of consolidation, such as reduction in energy and water consumption remain contested, and identified a number of subsequent studies to support this. Of these, the study by Myors et al. (2005) acknowledges its limitations by potentially sampling higher income locations, and the conclusion by Dey et al. (2007) that inner city locations demonstrate the highest levels of environmentally damaging consumption could equally be explained by their own findings that these consumption patterns are also typical of smaller and higher income households; households which now tend to concentrate in the inner cores of Australian cities. Other studies confirm correlations of higher consumption with income and education over the impact of urban form (Echenique et al., 2012; Grosvenor, 2013).

The main social benefits attributed to compact cities relate to improved health outcomes, equity and quality of life improvements associated with improved access to essential services, and improved opportunities for those on lower incomes. Ewing et al. (2014) find that more compact urban forms have the potential to reduce obesity and chronic diseases. This result is affirmed by a significant body of literature on this topic (as reviewed by Ewing and Hamidi (2015)). Other research has shown that residents who move into more compact urban forms tend to lose weight, but high weight individuals tend to self-select lower density urban forms thereby challenging the notion of environmentally dependent health results (Plantinga &

Bernell, 2007). There is less research in relation to access to services. Proximity to everyday services is typically improved in more dense urban environments (Kaido & Kwon (2008) cited in OECD, 2012, p. 62), although previous discussion on transport accessibility (Ewing & Cervero, 2010) is likely to be relevant to a more comprehensive understanding of this issue. Sprawling urban forms are often considered to provide more affordable housing however recent research has shown no correlation between sprawl and housing affordability except for very low income households for whom increased transportation costs of automobile dependency may still result in a worse outcome (Aurand, 2013). Improved affordability is often touted as a benefit of Australian compact city policies however Searle (2007) sees little evidence of urban consolidation reducing housing prices.

Although the debate around the benefits of compact urban forms continues, on the bulk of the evidence, it appears that when compared with sprawling settlement patterns compact urban form is correlated with changes in travel behaviour (away from automobile use), a reduction in greenhouse gas emissions (in the form of energy efficiencies), reduced traffic fatalities (but an increase in accidents), improved public health, reduced infrastructure costs, and improved economies in downtown areas (Ewing & Hamidi, 2015). Policy makers across the globe have apparently accepted this message, with policies for compact cities now ubiquitous in most OECD nations (OECD, 2012).

2.1.1. Compact activity centres and planning in Australia Australia’s urban development is often categorised into three primary stages: initial urbanisation, ‘the long boom’, and urban restructuring (Forster, 2004). Planning policies to foster the development of activity centres made their initial appearance during the “long boom” (approximately 1945-1970), with Sydney’s 1952 Cumberland County Council Plan and Melbourne’s Metropolitan Board of Works (MMBW) Plan of 1954 identifying centre locations with the intention of dispersing commercial and retail uses and limiting sprawl (Quirk, 2007; Searle, 2007). This period was formative for planning policy, with planning legislation becoming more formally institutionalised by most state governments (Forster, 2004; Sandercock, 2005). The plans developed during this period were in some ways not dissimilar to metropolitan plans today; large scale “masterplans” with a basis in the management of growth predicated on past trends, to be implemented via land use controls (Bunker, 2012; Sandercock, 2005). Although some plans included elements that attempted to control sprawling suburban development, such efforts were generally ineffective with “the private sector prov[ing to be] remarkably adept at influencing the decisions of planning agencies, either before plans were released or, when necessary, by pressing for changes to

29 published plans if those plans threatened their material interest”; an extension of the established “right” of property speculation (Forster, 2004, p. 22; Sandercock, 2005, p. 319).

Starting from the 1970s, the urban restructuring period continued sprawling suburban settlement as the primary form of housing to accommodate growth, although attached housing started to be developed in larger numbers in established suburbs in Brisbane, Sydney and Melbourne, particularly from 1990 onwards (Forster, 2004). Public dissatisfaction with the overly bureaucratic institutions of planning saw a reconceptualisation of planning (Freestone, 2012; Sandercock, 2005). This brought planning more formally into the sphere of political governance rather than through the previously established form of “blueprint” planning, whereby a bureaucracy of planning professionals attempted to coordinate service delivery (Gleeson & Low, 2000a, p. 72). This phase has been termed “social democratic managerialism” (SDM) by Gleeson and Low (2000a, pp. 74-76), a form of governance which saw increased political influence of planning bureaucracies and a more corporate style of management, combined with a focus on the equitable state provision of services to improve living standards and social justice. This approach was most enthusiastically embraced by South Australia’s Dunston government and through the Whitlam Federal Government’s Department of Urban and Regional Development, but at various times became an approach that was more or less adopted by most state governments (Gleeson & Low, 2000a; Parkin & Pugh, 1981; Sandercock, 2005).

The SDM movement was occurring during a time that coincided with the rise of neoliberal ideology which advocated for the replacement of centralised government planning with market based allocation of services (Gleeson & Low, 2000b). The ascendency of free-market ideology and failures of the SDM model to work within increasingly constrained budgets saw SDM enjoy a short life span before being replaced by a neo-liberal influenced, “corporate liberalism” style of governance (Gleeson & Low, 2000a, p. 92; Sandercock, 2005, pp. 325- 326). This “entrepreneurial turn” saw planning departments revert to the use of land use regulations to effect change, whilst other sectors of government sought private sector investment for mega-projects which, once realised, would override established land use controls and processes to “fast-track” the investment (Freestone, 2012; Sandercock, 2005, pp. 325-326). Planning became more project driven and increasingly saw planners involved in project management and urban design, as well as more involved in incorporating sustainability objectives into planning intents as part of the environmental awakening of the 1990s (Freestone, 2012). Regardless of the good intentions of planners to achieve sustainability and public interest objectives, some believe that “in practice, planning has been outsourced, privatised, markedsed [sic]” (Gleeson & Low, 2000b, p. 24), and to now be

30 primarily a marketing tool for attracting investment. Planners themselves have been required to adopt a somewhat confused “hybrid” role that attempts to balance both public and private interests (Steele, 2009).

Planning policy for activity centres also evolved during the urban restructuring phase. Initial centre policies of the 1950s were primarily focussed on the provision of goods and services, a rationale which over time expanded as a means to also reduce automobile dependence (Forster, 2004). The centres policy of Melbourne’s 1954 plan continued into the 1980s essentially unchanged (Quirk, 2007). The Melbourne approach attempted to develop multiple “district centres” for retail and office uses concentrated around public transportation nodes, however little attempt was made to assemble land for this purpose and retail and office developers disregarded the designated centres, instead preferring large sites on arterial roads convenient for access by automobile (Forster, 2004; McLoughlin, 1992; Quirk, 2007). With the exception of some centres in Sydney that were stimulated by the relocation of government departments, centres policy in other cities fared similarly, with disappointingly low concentrations of employment in designated centres (Forster, 2004). Issues related to sprawl also prompted changes to housing policy in the form of reduced development controls over new housing in established suburban areas (Dodson, 2012; Forster, 2004; Searle, 2007). This process drew on the concepts of the global “compact city” and “smart growth” movements and, in Australia, became known as “urban consolidation” (Dodson, 2012, p. 25). “These early deregulatory measures were accompanied by little overarching strategic planning guidance or attention to urban structure” and were therefore not usually integrated with plans to develop multi-centred urban forms (Dodson, 2012, p. 25). Urban consolidation policies began to generate considerable community opposition in the 1990s, particularly in higher income inner city neighbourhoods. In response to this opposition, urban consolidation policies began to receive more strategic direction which ultimately saw a convergence with activity centre policies and the emergence of the “containment, consolidation, and centres” metropolitan policies typical of today (Dodson, 2012, p. 26; Forster, 2006, p. 178; Searle, 2007). It should be noted that the term “urban consolidation” and the more internationally used terminology of “smart growth” and “compact city” often conflate land use policies to intensify populations in an urban core with policies of polycentricity that intend to do the same but within a network of regional sub-centres (OECD, 2012, p. 31). This study is primarily concerned with the latter type of policies.

2.2. Planning for compact activity centres in Brisbane

Planning for activity centres in Brisbane follows the overall narrative presented in the previous section; an increasing awareness of sustainability issues drove the establishment of

metropolitan level policy to create a network of compact activity centres characterised by higher density and more diverse residential uses in proximity to clusters of mixed uses providing employment and essential services. Starting with the “voluntary” policies of the Regional Framework for Growth Management of the 1990s, metropolitan policy became more formalised with the statutory planning of the South East Queensland Regional Plans that continue today. Over time, the intensity of development intended for the centres has increased, however the overall intent and implementation mechanisms for centres policy has been remarkably consistent for the past two decades. This section reviews available policy documents to detail the evolution of metropolitan activity centre policy during this period; from the incipient policies of the early 1990s, through to the first South East Queensland Regional Plan. The review also examines past attempts to monitor and evaluate the regional plans, and the latest iteration of the regional plan (SEQRP 2017). Although the latest regional plan maintains the centre policy of the past, it appears that the extent of the policy is now being questioned, and future policy may be more restrictive.

2.2.1. The beginning – the Regional Framework for Growth Management Compared to other major capital cities, Brisbane is unique in terms of local governance, where the majority of the broader conurbation’s population is within a single local government area. The size of the Brisbane local government area means that much of its land use planning had been occurring at a regional scale for many decades prior to establishment of more formalised regional planning that included the broader conurbation. Brisbane City Council had been undertaking planning for mixed use centres since the early 1980s, with a clear hierarchy of centre roles and functions (The State of Queensland, 1993b). Surrounding local governments (also large in area and population compared to urbanised local governments in other states) also typically made use of centre hierarchies, however there was limited consistency with other local governments in terms of how these hierarchies were classified (The State of Queensland, 1993b). There were also inconsistencies in the application of these centre hierarchies where large scale development was permitted contrary to the stated hierarchies, and a lack of “…regional or sub-regional context in which Local Authorities [could] attempt to resolve claims for centres to be the focus of subregional or regional level activity” (The State of Queensland, 1993b, p. 10). Cooperation between local governments was limited and “there was a culture of developer led growth [where] local councils, particularly those on the urban fringe, competed with each other to maximise their share of the growth” (Abbott, 2001, p. 115). A key rationale for the development of regional level plans was to correct such matters through the imperative of managing high regional population growth, thereby preventing sprawl and associated issues such as the loss of valuable agricultural and environmental areas (Abbott, 1995; Regional Planning Advisory Group, 1991). Previous attempts to do so had

32 achieved limited success through a lack of state government interest, and local government concerns about “threats to their autonomy” (Abbott, 1995, p. 135).

The first step towards more coordinated regional planning occurred through a state government led “community conference” in 1990 under the title of “SEQ 2001 – Framework for Managing Growth” (Abbott, 1995; Regional Planning Advisory Group, 1991). Conference participants were categorised into sectors comprised of federal government, state government, local government, social justice and environment organisations, business and industry organisations, professional bodies, and trade unions (Regional Planning Advisory Group, 1991). The proceedings of this conference do not have any direct reference to a regional activity centre network, however local government and professional sector input reveals support for more compact urban forms as a way of reducing issues associated with low density urban sprawl. Local government sought better coordination of new development with zoning that is more permissive for medium and high density housing; an objective supported by the community sector as a way to improve equality of access to services. The professional sector provided key principles that would subsequently form the basis for compact activity centres linked to sustainability principles, through support for planned new towns and satellite cities, increased housing choices and diversity, and a network based city focussed on expanding existing settlements and containing urban development to limit peripheral land release (Regional Planning Advisory Group, 1991).

The state government’s Regional Planning Advisory Group (RPAG - an organisation assembled to coordinate the delivery of the new regional level policy) led the conference and subsequent plan development through to 1994 (The State of Queensland, 1995). RPAG drew membership and input from a range of government, industry and community groups (Figure 1). The conference was used to inform a range of 14 “policy papers” which were subsequently

Figure 1 - Composition of the Regional Planning Advisory Group (Abbott 1995)

33 used in the preparation of “The Preferred Pattern”; a report that explored options in support of a recommended future regional urban form (Abbott, 1995, 2001; The State of Queensland, 1993c).

The policy papers and the “the preferred pattern” placed significant emphasis on the development of activity centres, with one of the policy papers specifically targeting centre objectives (The State of Queensland, 1993b). The Major Centres policy paper advocated for centres that provide a greater range of services to the surrounding community on the basis of equity of access, more efficient use of infrastructure, and efficiencies created from clustering businesses together (The State of Queensland, 1993b). The paper goes on to recommend a range of objectives, policies, and actions to achieve a network of activity centres. These include a focus on increasing employment opportunities, a mixture and diversity of uses, improving public transport links between centres, and improving amenity through the development of more activated spaces and improved streetscapes. The policies were to be implemented through updated regulatory land use plans that reflect these objectives, and the improved coordination of government service and infrastructure delivery. These methods remain as the primary implementation mechanisms still used in regional planning in SEQ today (The State of Queensland, 1993b, 2017c). Higher density residential uses were proposed in terms of improving mixed use outcomes and enhancing “the viability and attractiveness” of the centres (The State of Queensland, 1993b, p. 53). The sister policy paper, “Residential Choice and Efficiency” expanded on the rationale for higher density residential uses more in line with standard compact city rationales. More consolidated urban forms are considered important in achieving infrastructure efficiencies, improving access to employment and essential services, as well as reducing traffic generation (thus improving air quality and reducing fuel consumption) (The State of Queensland, 1993d). This policy paper specifically seeks to achieve “ecologically sustainable development” through a range of measures including more compact and dense residential uses. Facilitating the development of medium density housing and mixed land uses at centres and transport interchanges, was a key part of the document’s redevelopment policy (The State of Queensland, 1993d).

The initially proposed network of centres in the background papers (Figure 2) is remarkably similar to the activity centre network in use in the current iteration of the SEQ regional plan. These centres were to be comprised of “…a full range of shopping, administrative, cultural, community, entertainment, educational and recreational facilities” including department stores, supermarkets, speciality shops, restaurants, offices, medical centres, and a range of essential and community services. Services and trades, and light industrial activities, were to be accommodated in mixed business areas adjacent to centres. (The State of Queensland,

1993b, pp. 39, 79). For centres distant from the Brisbane CBD, medium density residential uses were proposed at a density of one dwelling per 300m2 of site area, which equates to approximately 33 dwellings per net hectare (The State of Queensland, 1993d).

Figure 2 - Proposed Centre Network from the 1993 Policy Papers (The State of Queensland, 1993b)

The use of these policy position papers is evident throughout the “Preferred Pattern” document, which shares similar overall principles and objectives. The Preferred Pattern used multi-criteria analysis of three different regional development scenarios (central concentration, coastal concentration, and coastal dispersed), to develop a preferred urban

35 form (The State of Queensland, 1993c). The resulting urban form was to be characterised by more compact urban development, with an increased rate of employment and community services growth in outer metro-Brisbane in three key strategic regional centres at Pine Rivers, Logan or Beenleigh, and Ipswich (The State of Queensland, 1993c). Five other Strategic Regional Centres were also identified which included Chermside and Mount Gravatt. These centres are mapped on the “preferred pattern of urban development” (Figure 3) (The State of Queensland, 1993a). Not mapped are series of “sub-regional centres”, which are instead defined by the size of the population catchment’s they serve (between 50,000 to 150,000 people). The plan seeks to “concentrate medium density housing around the major centres to provide opportunities for people to live within walking distance of a range of commercial and community facilities” (The State of Queensland, 1993c, p. 102). Other principles for centres included prioritising key centres over the CBD for government services, focussing employment in centres, and improving public transport. The policy called for a wide range of uses such as retail, health services, universities and other major employment generators. Large scale industrial uses were specifically excluded. The centre policy was described as being “critical” to achieving regional objectives such as air quality, energy conservation, infrastructure utilisation, travel costs, and improved accessibility to employment, shops, and services (The State of Queensland, 1993c, p. 100). The centres policy was to be achieved primarily through changing local government regulatory land use policies, as well as through direct government investment in centres, infrastructure development, and establishing joint ventures to assemble and development land (The State of Queensland, 1993c).

A separate, “Major Centres Inter-Departmental Working Group” published a report on RPAG’s centre policy, which sought to wind back the scope of government investment in centres in favour of an “evolutionary, market driven strategy” for implementation that

Figure 3 - "The Preferred Pattern" (The State of Queensland, 1993a)

provides local government with more “flexibility” (The State of Queensland, 1994a). The report voiced concerns about the level of resourcing and planning regulation that the preferred pattern potentially committed to major centres, where “the issue with respect to major centres implementation is not the extent of intervention per se but a pragmatic assessment of the costs, benefits and risks associated with the policy proposal at hand.” (The State of Queensland, 1994a, p. 22). Although overall supportive of compact activity centre policy objectives the report proposed a similar framework for plan implementation as the preferred pattern, but with the use of less-committal terminology and extended timeframes (Figure 4).

Figure 4 - A more "flexible" implementation process (The State of Queensland, 1994a)

A month prior to the release of this report, RPAG published the first iteration of the “Regional Framework for Growth Management” (RFGM), a final policy paper consolidating its previous planning, and completing the RPAG’s remit (Abbott, 2001). The centre policies in the 1994 RFGM were materially unchanged from those described in “The Preferred Pattern”, identifying the same key regional centres, noting the desire to develop medium density housing in proximity to centres, and justifying these changes as being important for economic, environmental, and quality of life reasons (The State of Queensland, 1994b). The plan noted that continued “sub-regional” planning was required and although the centre policies from the 1994 RFGM were supported in these plans, a number of local governments (including Brisbane City Council), disagreed with the proposed population distribution and densities (Abbott, 2001). The RPAG was replaced in mid-1994 by the Regional Coordination Committee (RCC), which now consisted only of local, state and federal government representatives with non-government roles being incorporated into an advice body known as the Regional Non-Government Sector Committee (RNGSC) (The State of Queensland, 1995, p. 6). The sub-regional plans were incorporated into the 1995 RFGM, which was endorsed by the RCC in September 1995, making it the first time in Australia that all three spheres of local government formally agreed to a metropolitan planning strategy (Abbott, 2001; The State of Queensland, 1995).

The 1995 RFGM maintained the same general principles for major centre development, however the plan reduced the number of nominated regional metropolitan centres to three; Beenleigh, Ipswich, and Caboolture. Other centres were included by definition only, which now featured a raised threshold for a major district centre of over 100,000 person catchment and 4,000 jobs (The State of Queensland, 1995). Several of the previously included centres meet these criteria such as Chermside and Mount Gravatt, however curiously the key regional centre of Beenleigh falls far short of even the major centre categorisation, with only 2,500 estimated jobs in 1996 (see section 5.3). A priority action of the RFGM was to undertake a future study to identify appropriate major centres (The State of Queensland, 1995). A 1996 scoping study identified 16 “candidate” centres that met the RFGM requirements for Major District Centres (Planning Workshop Australia, 1996). These centres included the majority of the nominated metropolitan activity centres in the subsequent SEQRP, except for Goodna, Wynnum, and Redcliffe (Figure 5). The primary implementation mechanisms for the centre policy continued to be planning based, with calls for improved coordination of infrastructure delivery and services, and new regulatory centre plans to accommodate increased

Figure 5 - Major District Centre Scoping Study Candidate Centres (Planning Workshop Australia, 1996)

employment, mixed uses, and high-quality urban design. However, implementation of the plan was temporarily suspended after a change of state government resulted in the plan being reviewed (Abbott, 2001). The review primarily changed a range of governance matters and introduced new economic principles, but did not make significant changes to the overall direction (The State of Queensland, 1996). The resulting amendment to the RFGM was completed in December.

The policy for activity centres continued essentially unchanged through the following two iterations of the RFGM (The State of Queensland, 1998, 2000). Both of these revised plans noted that local government planning schemes had now recognised the key metropolitan centres, and by 2001 “key centre development strategies” were either completed or underway in all key centres (SEQROC, 2001). The call for revised planning provisions that accommodate centre policy remained, although new terminology was introduced to focus changes on regulations that “unnecessarily restrict the market from providing choice in housing to meet demand and achieving higher densities in appropriate locations” (The State of Queensland, 1998, p. 52). Major district centres were not spatially identified in either of the plans, and actions to undertake studies to identify these centres remained. The 2000 RFGM was to be the final iteration of this series of plans, and noted that a comprehensive review of the RFGM was to commence after the publication of the 2001 census results, to culminate in a new plan by 2004 (The State of Queensland, 2000).

2.2.2. The regional plan becomes statutory – The South East Queensland Regional Plan True to this promise, the Queensland Government released a new draft regional plan in 2004 (The State of Queensland, 2004). The draft plan was released as a formal “planning instrument” under the Integrated Planning Act. For the first time, regional planning documents had statutory force for the consideration of development applications, as well for making and amending local land use regulations. With this new era, came a dearth of publicly available background reports and other policy documents which informed the development of draft regional plan (such as the policy papers released by RPAG as part of the development of the RFGM). The draft plan received considerable public interest with more than 8,000 submissions, many of which originated from property owners now effected by the plan’s urban growth boundary, preventing future urban development (The State of Queensland, 2005a). Submissions in relation to activity centres were generally supportive and primarily related to questions around the level of government investment (particularly in transport infrastructure) as well as issues of centre scale and expansion. Submitters also saw the plan’s implementation section as being weak, with insufficient detail especially around key responsibilities, and how the land use planning system alone would not achieve the planned outcomes (The State of Queensland, 2005a, p. 122).

The adopted South East Queensland Regional Plan was released in June 2005 and the consultation process resulted in only minor changes to centre policy. Just like its predecessor, the SEQRP gave significant prominence to the concept of consolidating future urban growth “around urban activity centres” (The State of Queensland, 2005c, p. 12). The plan provided “Regional Policies” which contains a number of “Desired Regional Outcomes” (DROs). Of

40 these, DRO8 specifically deals with compact centres, calling for “a compact and sustainable urban pattern of well-planned communities, supported by a network of accessible and convenient centres close to residential areas, employment locations, and the transport” (The State of Queensland, 2005c, p. 60). The DROs are then split into a number of principles, which include objectives to make more efficient uses of land by “focus[ing] higher density and mixed-use development in and around regional activity centres and public transport nodes and corridors” (The State of Queensland, 2005c, p. 65). Principle 8.6 is dedicated to creating a network of activity centres where employment and services will be focused in a network of “well-planned, vibrant and accessible regional activity centres”, with sub-principles calling for the definition of a network of centres, alignment with local government planning, encouraging major new developments within centres, supporting centres with transport infrastructure, and having detailed local plans for each centre (The State of Queensland, 2005c, p. 71). The principles also cover issues of proving “a range and mix of dwelling type, size, and locations” and “character and design” aspects to improve urban amenity (The State of Queensland, 2005c, p. 69). The plan justifies the principles in terms of their contribution to sustainable development.

These strategic objectives for compact activity centres are very similar in nature to those detailed in the RFGM plans from the 1990s. The main difference with the SEQRP relates to the number of activity centres spatially nominated in the plan. While the RFGM only spatially identified 3 key metropolitan centres with the remaining major centres being defined by their catchment sizes, the SEQRP maps all of the centres identified in the hierarchy (Figure 6). It is interesting to note however that the SEQRP centres are mostly the same centres as those initially identified in “The Preferred Plan” (Figure 2, p35).

Figure 6 - 2005 Activity Centre Network (The State of Queensland 2005c)

Although the terminology of the centre types has changed, their overall role and desired uses are consistent with the major centre descriptions in the RFGM. The proposed density for the centres had increased under the SEQRP, from the “medium” density development of 33 dwellings per hectare, to “higher” density development of 40 to 120d/ha (or greater) for principal activity centres, and 30 to 80 du/ha (or greater) for major activity centres. These residential densities were expressed in terms of “net” density, which the plan defined as the number of dwellings divided by the residential lot areas plus the area of local roads and parks (The State of Queensland, 2005c, p. 133). This definition is quite imprecise as it is difficult to know which parks or roads are to be included in the calculation.

Like the RFGM, implementation was primarily to occur through the regulatory land use planning system, with some government support in terms of investment in services and infrastructure. State government infrastructure investment aligned with the SEQRP policy was detailed in the South East Queensland Infrastructure Plan (SEQIP) (The State of Queensland, 2005b). Transport infrastructure was overwhelmingly the largest component (in monetary terms) of the proposed infrastructure. The Western corridor was allocated a large portion of this funding including a new train line to Springfield, and motorway upgrades linking Ipswich and nominated growth areas. Greater Brisbane received significant funding but only a few activity centres were directly affected (mostly in the form new busway links to Capalaba and Springwood, which as of writing have yet to be realised). Motorway upgrades were also proposed. Hospital upgrades were also considered in terms of compliance with activity centre policy, such as the upgrades to the hospitals at Ipswich and Chermside.

An update to the SEQRP 2005 was released in 2006 (The State of Queensland, 2006a), and in July 2009 was superseded by the SEQRP 2009 (The State of Queensland, 2009). This plan continued the centre policy from the previous plan mostly unchanged. Similar justifications against sustainability principles remained, however the plan included an additional justification for higher density residential development near centres in order to protect against oil vulnerability. Aspects related to design had been slightly expanded and included additional justifications for active frontages and mixed uses in order “to enable residents, business people and workers to meet and interact, build social capital, and create networking and business opportunities” (The State of Queensland, 2009, p. 100). The implementation section of the plan remained weak (a single page), and similar implementation mechanism were proposed. This time however, use of statutory plans to implement activity centre policy is expressly described as the “key delivery mechanism” (The State of Queensland, 2009, p. 9).

2.2.3. Monitoring and evaluation of the regional plans The call for monitoring and evaluation of the regional plans has been present since the initial RFGM days, where the plans called for a “review of demographic, environmental, and economic trends in SEQ” (The State of Queensland, 1995, p. 76). It has not been possible to find reports that detail this monitoring beyond the summaries provided in various iterations of the RFGM. The first published report on evaluating plan outcomes that could be accessed comes from 2006, through the Urban Development Monitoring Program (The State of Queensland, 2006b). This report focussed on dwelling activity, and how well residential growth aligned with regional planning objectives. Subsequent stages of the program were to investigate commercial and industrial development, and the supply of land. The program however did not progress further (at least in a format that released its results publicly), and it appears as though this monitoring ceased in 2007 (The State of Queensland, 2017a, p. 49), with no reports published beyond December 2006. The 2006 report described total new dwelling activity and targets, infill and redevelopment activity and targets, demolitions, lot approvals, production, lapsed, registration and totals, median house and unit/townhouse prices, and vacant land sales. The report only focussed on change over a two year period. It’s questionable whether the impacts of regional policy would be visible in such a short time period. The report went into some detail on residential dwelling change, matched these against broad scale dwelling targets for the region and various local government areas, and concluded that dwelling development was generally on track to meet the plan’s targets.

The urban monitoring reports appear to have been replaced by the more comprehensive State of The Region reports from 2008 (The State of Queensland, 2008). This report provided a relatively detailed description of the results of a range of indicators used to evaluate the conformance of the new regional plan. The actual timeframe varied, but it mostly measured changes between 2001 and 2006. Policies for the compact city and activity centres were primarily captured under DRO8 – Urban Development of the SEQRP 2005. The report gave a “green light” to most of the indicators in this section including urban form and employment in activity centres (The State of Queensland, 2008, p. 284). Again, this is a very short time frame for analysis of regional policy. The “urban structure” indicator measured the proportion of new residential within/out of the urban footprint and noted that this is working well with 94% of new dwellings in the footprint. The “urban form” indicator examined the extent of new residential achieved through infill and redevelopment of existing urban areas (The State of Queensland, 2008, p. 287). Over a three year period the report noted that the infill targets were being met, although the number of approvals were declining. These conclusions are based on ABS dwelling approval data sets. The “employment in regional activity centres” measured total employment in selected activity centres over the period of 2001 to 2006. The

44 measure includes the Brisbane CBD and the greenfield centres like Springfield which start from a low base. From the study area, Upper Mt Gravatt is highlighted as having high employment growth, but no other centre grew by more than 1000 jobs, and Springwood declined. The study didn’t track employment growth more generally and used the ABS journey to work data. This data has a range of issues in terms of the size of the destination zone areas and how variable these are across centres (see section 4.2.2). The study does not describe how they accounted for this, or what they used as boundaries for activity centres. The “housing mix” indicator used ABS dwelling approvals to examine the proportion of new dwelling approvals by type of dwelling (The State of Queensland, 2008, p. 294). This indicator was given a yellow light. It showed a year by year decline in the proportion of apartments, and an increase in the proportion of detached dwellings. The report noted significant variations between local government areas with “higher density dwellings types (i.e. attached houses and apartments) concentrated in the more urbanised local government areas” (The State of Queensland, 2008, p. 294). This was the only SoR report that could be found in publicly accessible records.

The next evaluation report that could be obtained was the Growth Management Program Annual Report. This report was not as comprehensive as the SoR Report but provided a detailed overview of aspects related to dwelling activity, dwelling supply and industrial development and land supply (The State of Queensland, 2010). Its scope was to examine progress of the SEQRP 2009 in terms of a range of planning objectives, including Outcome 8.1 for compact development. The report is a continuation of work from the Urban Development Monitoring Program and measured land supply in terms of “approved supply” (dwellings with an approved development application), “planned supply” (land that is suitable for development subject to a DA), and “emerging supply” (land that is being considered for future development but for which the planning system does not yet formally permit) (The State of Queensland, 2010, p. 18). The report noted that overall, dwelling supply was mostly in line with the requirements of the plan, and infill dwelling activity was ahead of projections. Unlike the SoR report, there was no direct analysis of the activity centre policies.

From 2012 on, the new Department of State Development Infrastructure and Planning reduced the amount of publicly accessible information available on the regional plans, and monitoring and supporting information sections were removed from the department’s webpage. The department presumably continued to undertake some monitoring and evaluation of the regional plan, however it has not been possible to find any updated versions of the monitoring and evaluation reports in the public records. Overall, the Queensland Government appears to have maintained sporadic interest in evaluating its regional plans (The

State of Queensland, 2017a, p. 49). The new SEQRP2017 however places substantially more detail and importance on evaluation and monitoring in future regional planning,

2.2.4. The current regional plan From 2012 the Queensland Government began promoting a new regional plan to be released in 2014. The first iteration of the new regional plan was aborted with a change of government, and preparations for a revised plan commenced under the new government. Plan development involved a number of consultations with the public and interest groups in the form of “community conversions”, “talk to a planner” sessions, a “youth summit”, and calls for submissions on a draft plan (The State of Queensland, 2017d). As part of the results of the initial consultation period the department provided a selection of “great ideas” provided by the community. The compact activity centre featured highly in the department’s conception of greatness, where they highlighted comments about higher density development around shopping centres, mixed use development, and more detail for regional activity centres:

“Regional activity centres development needs guidelines—The development around the Principal Regional Activity Centres (PRAC) need to be more structured and be provided with detailed guideline by the updated SEQRP. Currently the SEQRP only suggests that areas around PRAC have higher density. As part of this revision of the SEQRP guidelines for development around PRAC should be specified. For example, within 400m walking distance should be a minimum of eight storeys and within 800m walking distance should be a minimum of five storeys. There are examples of places being 800m from PRAC being 10 storey or 1.2km being six storey while other areas that are only 500m from the same PRAC are only low to medium density. This inconsistency should not be allowed to occur and if you are within walking distance (800m) it should be at least medium density.” (The State of Queensland, 2016a, p. 24).

These aspects were incorporated into the draft and adopted versions of the new regional plan through the inclusion of detailed (and substantially more intensive) residential development requirements for activity centres (Figure 7). Otherwise, the new regional plan continues the activity centre policy seen in past plans, with the same network of centres.

Figure 7 - Residential densities for activity centres in the SEQRP2017 (The State of Queensland 2016b)

The new plan also offered an expansion of development opportunities through new “infill corridors” between centres, a “missing middle” (redevelopment of middle suburbia where it meets certain criteria), and new “knowledge precincts”, industry areas, and “regional economic clusters”. Centre definitions are materially the same as past plans but have been adapted to include new nomenclature around “knowledge intensive business”, “compet[ing]in globally competitive markets” (The State of Queensland, 2016b, p. 54).

The implementation section of the report has been significantly expanded compared to previous plans, but the overall concept remains similar; implementation is to occur primarily through the development assessment process with the regional plan creating alignments with other government actions such as infrastructure planning and local government statutory plans. New actions include catalyst projects to demonstrate innovative ways of integrating residential and mixed uses, design manuals, and more planning. There is a significant emphasis on monitoring and evaluation. A new SEQ Growth Monitoring Program is to be established to review the implementation of regional plans. Of particular interest however is a future review of centres planning policy (The State of Queensland, 2016b, pp. 123-124). The department has maintained the existing centres network policy until this review is complete however the department now acknowledges significant issues with centre implementation:

“Over the last 10-year period, a number of centres have not reached their expected capacity to be mixed use, support higher density living or provide the social infrastructure required to meet the demands of the community. Anecdotal evidence suggests that the likelihood of some centres to ever be truly mixed use or have significant in-centre residential population is doubtful. Exacerbating the issue is the unsustainable number of centres currently in the network, which is likely diluting focus from critical planning outcomes and investment in priority infrastructure and services”. (The State of Queensland, 2017b, p. 23).

2.3. Conclusion

This chapter has provided background and context on the development of the sprawling urban form that is now typical in Australian cities, the emergence of a particular form of planning culture and governance, and how this contributed to the current compact active centre policies that seek to mitigate the impact of sprawl. Improving economic efficiency of the delivery of infrastructure initially drove attempts to contain sprawling suburban development however this reasoning was later broadened to include a host of purported sustainability benefits to justify the adoption of policies to prompt a higher density, more compact urban form, in line with the compact city movement. Whether a compact city produces such benefits continues to be debated, however the weight of research tends to support some of its stated benefits. Compact city policies have been adopted across all major Australian cities and are now typical throughout OECD nations. Australian compact city policies rely heavily on the development of a network of compact activity centres to achieve the sustainability benefits that are supposedly inherent in such a city form. These policies have now been a key feature of metropolitan policy in all major Australian cities for the past two decades. Metropolitan Brisbane is no exception. A review of more than two decades of policy documents revealed that regional level planning has had a consistent vision for the establishment of a network of compact activity centres for this period. Although there have been some changes over time in terms of the intensity of desired development, the overall policy has shared similar principles since its inception. These principles seek the development of centres characterised by:

• Higher density residential uses in and near centres;

• A greater diversity of housing types; and

• Mixed clusters of uses that generate employment and provide localised services.

The rationale for these policies has been based primarily on the supposed sustainability benefits of more compact urban forms; an idea that has increasingly become part of mainstream planning thought since the 1990s. The multi-nodal expression of Australian compact city policies is seen as an evolution from previous centre based planning approaches initially established in the 1950s. The existing state led attempts at monitoring their own activity centre policies have so far been sporadic, made use of insufficient time periods, and do not detail how critical methodological issues were addressed. The next chapter reviews theories of plan evaluation, other evaluations of activity centre policy, and how these relate to greater Brisbane.

3. Plan evaluation theory, results, and methods

This chapter reviews the theory, practice and methods of plan evaluation and identifies key gaps in existing research. Literature on theories of plan evaluation are firstly presented before examining existing research in relation to compact activity centre policy implementation. Finally, the methods used in plan evaluation are considered in relation to methods commonly used to measure the compact city.

Plan evaluation theories are primarily split by approaches based on conceptions and definitions of planning itself. These approaches differ based on whether they evaluate plans in terms of how well the plans have been used by decision makers (plan performance), or the more positivist approach which evaluates plans on the basis of whether its objectives are reflected in changes to the physical realm (plan conformance). Although there are studies utilising each approach, as well as combined approaches, evaluation scholars continue to claim that there is insufficient research of plan implementation, particularly of policies that seek to change urban forms, such as compact city policies. The literature also calls for more research into the development of suitable methods for plan evaluation.

Existing empirical research that measures progress towards implementing compact activity centres in Australia, is entirely conformance based. The results of this research show that activity centre policy has had mixed conformance, with particularly slow changes in the existing centres of the middle and outer suburbs (Bunker, 2014; Chhetri, et al., 2013; Newton & Glackin, 2014). Numerous factors have been proposed to explain the slow progress of implementation of activity centre policy. These studies challenge the policy’s feasibility based on factors that include property economics, consumer preferences and demographic patterns of employment distribution, transport accessibility, and flawed plan making and governance. These explanations are typically considered in terms of the local influences of place, economics, and politics. The existing plan implementation studies however rarely make use of conceptions of theories of plan evaluation, and there has yet to be a study on this topic that uses a framework of both conformance and performance to specifically evaluate activity centre policy.

Much of the existing analysis of progress towards developing compact activity centres relies on data that is now more than 10 years old and/or covers an insufficient extent of time to allow for planned affects to materialise. There are also a range of data constraints that pose severe limitations on past studies. The only evaluation of the greater Brisbane area to date for example, is subject to all these issues and, as discussed below, therefore presents unreliable conclusions. The remaining Australian empirical research is almost exclusively focussed on

49 the case of Melbourne. An evaluation of activity centre policy using current data, and which compares performance and conformance with further explanatory factors, has yet to be undertaken.

3.1. Plan Evaluation

The widely cited critique of planning by Wildavsky (1973) was a clear challenge to the purpose and efficacy of the formalised act of planning. Central to his argument was the idea that “promise must be dignified by performance” (Wildavsky, 1973, p. 129). Such calls have led to renewed interest in the evaluation of planning, and the theories and methods that are most appropriate to do so (Alexander & Faludi, 1989; Oliveira & Pinho, 2010a; Talen, 1997). Theories of planning and plan evaluation have been influenced through the various conceptualisations and definitions of planning, however even though evaluation has been considered as a vital component of the planning process for more than four decades (Lichfield et al., 1975), evaluation still “remains relatively unexplored and under-used” (Guyadeen & Seasons, 2018, p. 107).

Before proceeding it is necessary to draw a distinction between evaluation in planning and evaluation of planning. This distinction is primarily one of timing. Evaluation in planning, commonly referred to as “a priori” evaluation, relates to attempts to evaluate alternative strategies in line with projected future outcomes so as to select the optimal strategy, and therefore occurs prior to plan implementation (Alexander, 2006, pp. 7-8). Evaluation of planning, or “ex post facto” evaluation, is concerned with the evaluation of the outcomes of planning so as to learn from experience (Alexander, 2006, pp. 7-8). Alexander (2006) also considers an “evaluation in progress” where implementation is monitored for conformance with key objectives or indicators. Where plans are frequently updated or renewed but maintain enduring concepts, such as the compact city as discussed earlier, there may not always be a clear distinction between ex post and in progress situations. This study is primarily concerned with the evaluation of planning. As such, it is focussed on evaluating plan implementation. This approach is necessary to better understand the effects of planning and is fundamental to address critiques such as those put forward by Wildavsky (1973). As Alexander and Faludi (1989, p. 127) state, “if planning is to have any credibility as a discipline or a profession, evaluation criteria must enable a real judgment of planning effectiveness: good planning must be distinguishable from bad.” Such a notion seems obvious, yet there are noticeably fewer studies relating to ex post and in progress evaluations of planning (Oliveira & Pinho, 2010a, p. 347).

Studies on plan evaluation can be further categorised based on the conceptual approach used as part of evaluation. A key difference in approach relates to studies that evaluate plan quality compared to studies that evaluate plan implementation (Connell & Daoust-Filiatrault, 2018; Sciara, 2015). Studies of plan quality are plentiful and typically make use of content analysis and surveys of the makers and users of the plans (Lyles & Stevens, 2014). In contrast, the evaluation of plan implementation continues to be widely considered as underdeveloped (Guyadeen & Seasons, 2018; Lyles et al., 2015; Oliveira & Pinho, 2010a; Sciara, 2015). Seasons (2003) explains that this lack of commitment to evaluation occurs due to a lack of resources, poorly developed methods with a lack of suitable indicators, difficulties linking cause with effect, and political and organisational cultures.

Studies of plan implementation differ based on how planning is conceptualised resulting in something of a dichotomy in the approach to plan evaluation. This has come to be popularly described as evaluating plans on the basis of “performance” versus “conformance”; two fundamentally different concepts of how “successful” plan implementation ought to be defined (Guyadeen & Seasons, 2016; Laurian et al., 2010; Loh, 2011; Oliveira & Pinho, 2009; Oliveira & Pinho, 2010a; Talen, 1997). These terms appear to have originated in scholarship related to business program evaluation in the 1960s (see Anthony, 1965), and continue to define plan implementation evaluation today (Guyadeen & Seasons, 2018). Plan performance evaluation is derived from the position that it is unrealistic for plans (particularly strategic level plans) to be considered as blueprints for a future end state, and that an evaluation of the success of a plan ought therefore not be focussed on its material effects (Faludi, 2000). Instead, a plan is considered successful if it is invoked by decision makers when considering a relevant issue. As planners rarely have the means to implement the plans they create, it is reasoned that plan success is therefore dependent on the decisions made by those who do have the means to implement plan objectives (Faludi, 2000). Under this view, whether a decision conforms to the plan or results in the material effects described by the plan, is not relevant to successful implementation; implementation is instead considered to be a success if the plan is used (Alexander & Faludi, 1989; Faludi, 2000). This is considered to be especially applicable for broader scale regional level plans, where the intended outcomes are more strategic. Faludi (2006) illustrates this concept in a study that examines the patchy conformance of the European Spatial Development Perspective in terms of how it is used by decision makers in various EU member states. Although the plan was not always followed, Faludi (2006) concludes the strategy was successful as it was consistently referenced in decisions by the countries involved and therefore its ideas had been “absorbed”. Conversely, Altes (2006) describes a situation where a plan achieved the initially intended results however it was not used in making the decisions necessary to adapt to changing circumstances. The performance

51 based approach to plan evaluation was therefore justified on the basis that although the plan achieved results in conformance with the initial plan goals, the plan was unsuccessful as these results were no longer those required (Altes, 2006).

Plan conformance evaluation is typically considered as being derived from a positivist approach that seeks to directly connect planned objectives to changes in the physical world (Alexander & Faludi, 1989; Loh, 2011; Oliveira & Pinho, 2009; Talen, 1997). Conformance evaluation sees a plan as a guide for future urban development and the key concern for implementation as being “the link between the plan and outcomes on the ground” (Oliveira & Pinho, 2010b). Alexander and Faludi (1989) view the critique of planning by Wildavsky (1973) as requiring absolute plan conformance for a plan to be successful. Using a strict conformance approach, a plan which does not achieve its objectives would be considered a failure. This is of course a rather extreme position which Alexander and Faludi (1989) critique as being incapable of incorporating the necessary uncertainties associated with the planning process, accounting for partial implementation successes, or beneficial outcomes enabled by a plan even if not directly envisioned by it. However, provided uncertainty is recognised, this does not remove the value of evaluating conformance. In most instances plans are devised to be effectuating devices and those involved in their development will have an interest and expectation in its physical implementation. Talen (1996, p. 90) states that decision makers are let “off the hook” when plan conformance is not evaluated, and discounting plan conformance evaluation “on the basis of the uncertainty factor can be seen as evaluation avoidance”. It is argued that uncertainty and flexibility can be considered as part of the interpretation of the results of conformance evaluation, and that issues associated with multicausality should not inhibit studies of conformance provided the investigation is focussed on the degree to which the plan achieved its goals (Loh, 2011; Talen, 1997).The literature clearly indicates that there are benefits to evaluating plan implementation from both performance and conformance perspectives. Even from the early stages of the development of this literature scholars have been advocating for the inclusion of both in evaluation methodologies, such as the Policy Plan Implementation Process (PPIP) method proposed by Alexander and Faludi (1989). Some combined approaches to plan evaluation include studies examining plan conformance and plan quality (Brody & Highfield, 2005; Burby, 2003) and, particularly more recently, a number of studies have incorporated measures of both plan performance and plan conformance (Altes, 2006; Berke et al., 2006; Feitelson et al., 2017; Lyles, et al., 2015; Zhong et al., 2014). These studies indicate that plan performance does not necessarily result in plan conformance and vice versa (Altes, 2006; Feitelson, et al., 2017).

The subject of this study is a regional land use policy that seeks to reshape the urban form to develop a series of compact activity centres. The activity of land use planning naturally involves proposals to alter the physical nature of a geographic area. For a land use plan to be implemented in terms of conformance, at some point decisions will need to be made to mobilise the necessary resources to create a material difference in the world. The nature of these decisions will also have an impact on what is physically delivered. Evaluation can therefore be considered in terms of a matrix that plots plan failure and success on a combined spectrum of both performance and conformance (Table 2).

Table 2 – Matrix of conformance vs. performance in evaluative approaches (author)

Considered in these terms, an evaluation of both performance and conformance is necessary in order to explain plan implementation successes or failures.

3.2. Evaluation of the implementation of compact activity centres

The research to date indicates that slow (or at best mixed) progress has been made after 20 years of compact city policy in Australia; particularly in relation to developing a more poly- nodal settlement pattern in the middle and outer rings of Australia’s largest cities (Bryant, 2013; Bunker, 2014; Chhetri, et al., 2013; Dodson, 2010; Newton & Glackin, 2014; Searle, 2010; Woodcock et al., 2011). However, this existing research does not include considerations of plan evaluation theory and is entirely conformance based. It therefore pays little attention to how the regional scale policy has been used by subordinate policy makers to develop planning regulations or otherwise apply the policy. Without such considerations of performance, it is difficult to determine whether the lack of conformance is due to the poor use or application of the policy, or whether other reasons explain the policy failure. The existing research is also almost entirely focussed on the case of Melbourne, and typically uses data that is now more than a decade old. The following sections review the existing research and the current explanations for this lack of conformance.

3.2.1. Existing evaluations of compact activity centres The most methodologically consistent analysis of the performance of Australian compact city policies comes from the Australian Government’s Bureau of Infrastructure, Transport and Regional Economics’ (BITRE) series of reports on “Population Growth, Jobs Growth, and Commuting Flows”. The BITRE reports specifically investigate the performance of activity centre policy and are the only comprehensive evaluation of activity centre policy for the Brisbane area that could be found. Unlike the studies from other centres, the BITRE reports tend to give an optimistic tone of progress towards the compact city noting that “significant density gains occurred in SEQ’s regional activity centres between 2001 and 2006” (BITRE, 2013, p. 107). Using the same reports however, Bunker (2014) identified some, but limited, progress towards compact city objectives whereby they have been effective in some larger inner centres but “…it has not generally happened as planned in other locations” (Bunker, 2014, p. 453). For example, the BITRE report for South East Queensland highlights that residential growth in activity centres has been greater than in other locations across the region (BITRE, 2013, p. 102). However, 57% of this activity centre population growth occurred in the Brisbane CBD and 3 other centres that were formed as ‘masterplanned’ greenfield developments. These locations are fundamentally different to the infill development that is necessary in the typical centres proposed by the policy. These more typical centres exhibit far more modest levels of growth, more in line with the regional average (BITRE, 2013). The BITRE reports also make use of 2001 and 2006 census data for their analysis of population and housing growth; data which is now ten years old and which presents challenges in terms

54 of use for comparative purposes (see section 4.2.2). Chhetri, et al. (2013) made use of property cadastral data from the same time period in their analysis of designated activity centres under the Melbourne 2030 plan. Their analysis found no statistically significant correlation between population densities and the plan’s designated activity centres (Chhetri, et al., 2013). Newton and Glackin (2014) had similar results when using more up to date data. They found that although infill housing development is occurring in suburban Melbourne, it is occurring in an ad hoc manner whereby only a relatively small number of infill dwellings (14% of the total number of infill dwellings) can be attributed to policies to cluster development in and near nominated activity centres (Newton & Glackin, 2014, p. 133). Phan, et al. (2009) also used changes to cadastral and address data to examine residential intensification in Melbourne. They showed that only a small proportion of residential intensification was occurring in proximity to activity centres, however this study only investigated a single outer-suburban local government area. These studies have all used relatively short time frames of less than 10 years.

These results align with the predictions of the early critics of the compact city concept who challenged the feasibility of the implementation of large scale alterations to the urban form (Breheny, 1997; Gordon & Richardson, 1997; Troy, 1996; Williams, 1999). Difficulties in establishing more compact urban forms has not been limited to Australia. In their recent review of the compact city concept, Ewing and Hamidi (2015) identified a number of studies that suggest progress towards implementation has been difficult and mixed. Of these, a study by Filion (2009) of Toronto’s long standing policy towards the development of compact “nodes”, a policy similar to the activity centre policies enacted in Australia, showed that the nodes had failed to attract the expected commercial and retail development for the past two decades and did not significantly contribute to improved pedestrian and public transport patronage.

It is important to distinguish between urban consolidation of general nature and compact activity centres. Considerable progress has been made towards urban consolidation of a general nature. In South East Queensland for example, 50% of new dwelling approvals since 2011 have been for attached dwellings (The State of Queensland, 2016b). The feasibility of urban consolidation in middle and outer suburbs in a greyfield setting has also been comprehensively analysed and these suburban settings are seen as economically suitable for small scale urban redevelopment (Newton & Glackin, 2014; Newton et al., 2011). However, compact activity centre policies in Australia seek to do more than simply increase the supply of housing in existing urban areas. The attempted shift towards a more poly-nodal settlement pattern is at the core of Australian attempts to establish more compact cities and to achieve

the sustainability objectives that are purportedly inherent with this city form. Birrell, et al. (2005) have provided a detailed critique of the Melbourne 2030 metropolitan plan’s activity centre policies which included a number of factors which they believed would prevent the plan from being implemented (discussed further in section 3.2.2). However, this analysis was primarily predictive in nature and is now more than 10 years old. Coffee, et al. (2016) use data from the 1981 to 2011 censuses to compare densities across cities, however their ring based approach examines settlement pattern more broadly and does not consider activity centres. An up to date and comprehensive analysis of the progress towards the development of compact activity centres, particularly outside Melbourne and over a longer time period than 10 years has yet to be undertaken.

3.2.2. Factors affecting the implementation of activity centre policies The feasibility of urban consolidation policies has been challenged on a number of grounds including consumer preferences and demographic patterns of employment distribution (Birrell, et al., 2005; Troy, 1996), property economics (Bryant, 2013; O'Connor & Healy, 2004; Rowley & Phibbs, 2012; Searle, 2004, 2010), and transport accessibility (Dodson, 2010). Evidence from other international examples tends to support these explanations where market forces, political and institutional commitment, and demographic shifts have been shown to hinder the implementation of consolidation policies (Brewer & Grant, 2015).

The economics of the property market are a recurrent theme in explanations for why urban consolidation policies may be unfeasible. As the role of supplying higher density housing and new commercial uses falls to private property developers, it stands to reason that consolidation will not occur where market conditions do not permit the sale of these developments with sufficient profit margins (Bryant, 2013). Research from Canada has similarly shown that market demands are a key factor in urban consolidation, regardless of political will (Grant, 2009). Newton and Glackin (2014) use a “redevelopment potential index” (RPI) to identify locations in Melbourne where the economic conditions are most favourable to housing redevelopment. Their analysis of Melbourne demonstrated that greyfield development (i.e. the redevelopment of existing housing stock) provided twice the number of new dwellings compared to brownfield development, and that the middle ring suburbs have the greatest potential for increased housing densities (Newton & Glackin, 2014, pp. 131, 137). The densities produced by this type of residential infill development (which typically consists of small scale developments that convert one dwelling into 2 or 4 dwellings (Newton & Glackin, 2014, p. 131) are unlikely to produce the level of vitality necessary to meet some of the normative objectives of the compact city (Jacobs, 1993, pp. 275-276). Birrell, et al. (2005) similarly suggest that only centres with sufficient services to “draw demand from a

56 surrounding regional economy” are likely to develop as intended by urban consolidation policies.

Dodson (2010) also blames economic conditions for a lack of consolidation in middle and outer suburban areas, noting these areas in Australian suburbs typically lack the “nodal concentration” and public transport links necessary to support higher density development. This represents both market and government failures resulting in a “catch-22” situation where density is unlikely to proceed without improved high frequency public services, but such services are not considered to be viable without sufficient land-use intensity (Dodson, 2010). Dodson (2010, p. 497) also argues that for public transport to be successful in achieving consolidation objectives it will need to account for the “dispersed travel patterns” of residents in middle and outer suburbia. Mees (2009) however questions the role of density as a means to improving public transport mode share, a view confirmed by Ewing and Cervero (2010) in the context of the United States. The US research instead linked public transport accessibility with increased mode share. Accessibility however does not appear to translate into improved urban consolidation with research from Melbourne showing little relationship between public transport accessibility (outside of inner city areas) and the development of infill housing (Newton & Glackin, 2014; Phan, et al., 2009). The literature therefore suggests that public transport provision alone is not sufficient to ensure the implementation of urban consolidation policies.

An analysis of Melbourne’s metropolitan plan for urban consolidation showed that the majority of designated activity centres were located outside areas with high levels of employment opportunities, and that the types of employment envisioned for activity centres is most likely to occur in proximity to the CBD (Birrell, et al., 2005). Birrell, et al. (2005) also noted that few activity centres were located in areas that demonstrated high levels of employment self-containment, and that these aspects will limit the implementation of the activity centre policies. Recent evidence from Melbourne tends to confirm the prediction of Birrell et al. whereby activity centre policies were not correlated with a growth in employment clustering (Day, et al., 2015).

Woodcock, et al. (2011) contest that it is the nature of a “performance based” planning system itself that prevents consolidation through its encouragement of property speculation, and the subsequent approval of too many developments. Performance based planning is an approach commonly used in Australian development control where decisions are made based on “flexible” performance outcome statements rather than highly prescriptive controls such as the number of storeys of a building, or its distance from a road. It is reasoned that this flexibility enables too many approvals, which subsequently reduce the ability of any single

57 development to be financed and constructed. Instead a planning system that facilitates consolidation within desired activity centres, while more comprehensively restricting development outside the nominated centres, may improve the delivery of consolidation outcomes (Woodcock, et al., 2011). The neoliberal nature of Australia’s planning system has also been criticised for its exacerbation of housing affordability issues (Sharam et al., 2015) which has implications for consolidation objectives that intend to improve equity.

Searle (2010), using South East Queensland as an example, criticises the land use plans themselves on grounds that they do not allocate sufficient activity centres in inner city areas where the demand for high density development is greatest, while Birrell, et al. (2005) contend that the centres are too poorly defined both spatially and conceptually. Further, the planning system primarily focuses on rezoning land and does not have the resources or powers to bring together sufficient numbers of lots to significantly change land uses in accordance with the active centre policies (Birrell, et al., 2005). These sentiments are echoed by a number of other authors who believe the activity centre policies lack the required methods to be successfully implemented (Forster, 2006; Gleeson, 2012; Gleeson et al., 2012; O'Connor, 2003). Critiques of this nature are not new, with McLoughlin (1992) concluding that previous attempts of planning interventions to positively shape the urban form were equally as unrealistic and had not properly considered implementation. The governance structures of this separate realm are also proposed as an explanation for the lack of efficacy of activity centre policies, with the fragmented institutional responsibilities of the various local governments and state government departments lacking the required coordination to implement metropolitan level policy (Gleeson, et al., 2012). A supposed erosion of core knowledge of the spatial and social sciences in the planning profession is also seen as exacerbating this situation and allowing for the creation of plans that “resemble glossy marketing documents rather than serious exercises in metropolitan analysis” (Gleeson, et al., 2012, p. 120).

3.3. Methods of plan evaluation

This section examines the methods typically used in plan evaluation, and their suitability for use in the evaluation of compact activity centre policy. Methods related to the evaluation of plan implementation from performance and conformance perspectives are firstly reviewed in a general sense. Following this, methods used in past studies of the evaluation of compact city policies are critically analysed to identify a direction for future research that aligns with theories of plan evaluation, and which overcome the identified data constraints.

3.3.1. Methods to evaluate performance and conformance The field of plan implementation evaluation continues to search for appropriate methodologies to evaluate plans. This is perhaps unsurprising considering the wide scope of matters covered by comprehensive plans that continue to be used in planning practice. Methods need to be adapted to suit the scale, context, and issue of interest, as well as conform to limitations of data availability. What is apparent however is that holistic implementation evaluation methodologies (such as the PPIP by Alexander and Faludi (1989)), though theoretically robust, have had limited practical application beyond use by their authors. Oliveira and Pinho (2010a, p. 351) note that even though widely cited, the PPIP has not been used in the two decades since it was published. These authors instead propose an alternative general evaluation methodology, the Plan Process Results (PPR) method (Oliveira & Pinho, 2009, 2010b). However, this method appears to have suffered a similar fate and no record of its use could be found. Loh (2011, pp. 272-273) notes that although such “exhaustive” methods are thorough, they produce “immensely detailed” results which prove difficult to interpret. Evaluation is also poorly utilised in practice even though practitioners acknowledge its importance, where a range of barriers to undertaking evaluation have been identified including a lack of resources, the difficulty in linking goals and outcomes, a lack of political interest, political aversion, and change adverse organisational cultures (Laurian, et al., 2010; Seasons, 2003).

Understanding whether the plan was used as intended when making planning decisions is considered an important component of plan evaluation as it provides richness to explanations of why or why not a plan was implemented. Under the performance view, a plan is considered as a “message” to decisions makers and the plan is successful when it is invoked in decision making (Faludi & Altes, 1994). Faludi and Altes (1994, pp. 414-415) propose a three step research design to evaluate planning performance:

1. “Identify the decisions on which the plan should have had a bearing.”

2. “Identify the commitments which decisions carry, together with the arenas for, and the critics of, their justification.”

3. Identify if the “plan has… helped in shaping the codes used in justifying subsequent decisions, and [whether] this improved the quality of the justification of decisions in terms of taking account of the wider field of choice.”

Once the appropriate decisions have been identified, they then need to be evaluated in terms of how the plan was used in support of the decision. One method of undertaking such an examination is to simply ask decision makers and other actors involved in the decision

59 making. Previous studies of this nature have typically used surveys (Feitelson, et al., 2017; Lyles, et al., 2015) or interviews (Oliveira & Pinho, 2010b). Alternatively, subordinate plans and other documents can be examined to determine if decisions accord with planned intentions (Altes, 2006; Faludi, 2006; Zhong, et al., 2014).

Existing research, particularly research investigating plan conformance tends to focus on specific components of a plan and select methods that best suit that component and the available data. For plan conformance, quantitative approaches are typical and range from relatively simple comparisons between measures of physical or permitted change (Alfasi et al., 2012; Chapin et al., 2008; Gennaio et al., 2009; Laurian et al., 2004; Loh, 2011), to studies making use of statistical correlations at a variety of complexities (Alterman & Hill, 1978; Berke, et al., 2006; Brody & Highfield, 2005; Brody et al., 2006; Burby, 2003; Feitelson, et al., 2017; Padeiro, 2016; Talen, 1996).

Evaluation of conformance often further differs based on the subject of its measurement. The two main approaches involve either the conformance between plan and permits, or the plan and observable changes in the physical world. An example of the former is the “Plan Implementation Evaluation” (PIE) method, which looks for evidence of links between a plan’s objectives, and the permits granted through the regulatory application of the plan (Laurian, et al., 2004). This method allows for a large range of potential planning issues to be evaluated and has been applied to topics as diverse as stormwater management, and urban amenity (Berke, et al., 2006; Laurian, et al., 2004). The examination of permits allows for the inclusion of highly localised issues, however it does not account for situations where the physical outcomes do not align with the permit or where permits are not enacted such as when they are obtained for speculative purposes (Woodcock, et al., 2011). Others combine measures of permit conformance with actual changes to land use (Brody & Highfield, 2005; Brody, et al., 2006; Zhong, et al., 2014). Brody, et al. (2006) examine the clustering of permits in relation to planned intentions to show that in areas with high development demands, authorities are more likely to issues permits contrary to their plans. However, these studies have so far been restricted to relatively broad scales (national or state level) thereby potentially missing more localised changes to the urban form.

More commonly, physical changes to land use are compared to plan intentions. In one of the earliest examples of this kind of study, Alterman and Hill (1978) used a grid overlay to assign whether land uses conformed or didn’t conform to a land use plan. They then statistically correlated conformance against a range of potential explanatory factors. Chapin, et al. (2008) used a more sophisticated, parcel based approach to observe whether changed residential land uses aligned with new policies to protect coastal areas from hurricane impacts. Their study

examined land uses both before and after the introduction of the policies; an approach, they argue, that had rarely been used before and which is necessary to observe the impacts of a plan. They note that at the time of their study, the research by Alterman and Hill (1978) remained one of the few investigations that measured actual land use changes to assess the implementation of land use policies. Since then, a number of other studies of land use implementation have been undertaken (Alfasi, et al., 2012; Feitelson, et al., 2017; Gennaio, et al., 2009; Loh, 2011; Padeiro, 2016; Zhong, et al., 2014). With the exception of Gennaio, et al. (2009), these studies tend to make use of relatively coarse categorisations of observed and/or planned land uses. Loh (2011) makes a compelling case that not all non-conformance is equal and develops helpful categorisation between non-conformance in terms of areas that have a different actual use to the planned use but which have yet to develop, versus uses that have developed contrary to the intent of a plan.

Oliveira and Pinho (2010a) link plan implementation evaluation with changes to planning practice where plans that include issues of urban form and morphology are common. They observe that plans are increasingly focussing on sustainability issues with associated strategies for more compact cities, urban design, and design guidelines, to shape city forms. However, they also note that the realisation of these urban forms and their supposed benefits “…is still more in the realm of beliefs than in theoretical arguments confirmed by practice” (Oliveira & Pinho, 2010a, p. 357). As such, there is a need for empirical evaluation of the practical application of urban form policies, such as policies attempting to implement the compact city (Oliveira & Pinho, 2010a).

Overall, the literature reveals surprisingly few studies that empirically investigate plan implementation from either performance or conformance perspectives. The evaluation of plan implementation is also a field in need of additional methods which are suitable to the evaluation of specific policy objectives such as the compact city.

3.3.2. Methods to measure compactness and their suitability in measuring change in activity centres The measurement of compactness itself is not novel; numerous attempts have been made both in Australia and abroad. For example, the OECD (2012) have developed a range of indicators to measure urban compactness that are based on measures of population density, “proximate development patterns”, access and use of public transport, and accessibility to local jobs and services. There has also been extensive research into methodologies for measuring urban sprawl which are also of use here as sprawl can be considered the inverse of compactness (Ewing & Hamidi, 2014). Ewing and Hamidi (2015) have developed an index that incorporates much of the research related to the measure of sprawl. Similar to the OECD’s

61 methods, this index makes use of indicators relating to population density, activity centring, and accessibility to employment, but uses a measure of street accessibility in place of public transport indicators (Ewing & Hamidi, 2015). A similar range of indicators was used by Yang (2008) in a comparative study of compactness in two cities. Burton (2002) makes use of an extensive array of compactness indicators in an article that has proven to be seminal in evaluations of the compact city. This research indicates that measures of density and mixed use are important to identify compactness (Burton, 2002).

A review of research measuring urban compactness/sprawl shows that the indicators used typically fall within the following categories:

• Dwelling and population density based indicators (Abdullahi et al., 2014; Burton, 2002; Chhetri, et al., 2013; Ewing & Hamidi, 2014; Lin & Yang, 2006; Newton & Glackin, 2014; Song & Knaap, 2004; Stathakis & Tsilimigkas, 2014; Yang, 2008)

• Mixed use and accessibility indicators (Abdullahi, et al., 2014; Burton, 2002; Lin & Yang, 2006; Song & Knaap, 2004; Stathakis & Tsilimigkas, 2014; Yang, 2008)

• Street network indicators (Ewing & Hamidi, 2014; Song & Quercia, 2008; Yang, 2008)

These groups of measures are aligned with common conceptions of the compact city, such as the OECD’s mixture of density, proximity, and accessibility (OECD, 2012). As such, there is also a similarity with Brisbane’s metropolitan policy objectives for compact activity centres. As discussed in section 2.2, these policies aim to create centres that are characterised by higher residential densities, a greater diversity of residential use types, and mixed clusters of employment generating uses that provide services to the local and broader community. It is important that indicators used in the evaluation of policy directly relate to the objectives of the plan. For example, although indicators related to street networks have been demonstrated to be good indicators of compact urban forms (Ewing & Hamidi, 2014), they do not directly relate to the planning objectives sought by regional planning in SEQ. They are also less suitable for measuring change in compactness over time in existing urban areas, as street networks are highly resilient to change, and there has been very little (if any) change to the street network of any of greater Brisbane’s centres (excluding the greenfield centres).

Issues of scale are also important considerations. Most of the reviewed indicators have been designed to examine entire metropolitan areas. However, the poly-nodal settlement pattern at the core of Australian attempts to establish more compact cities calls for the longitudinal analysis of discrete parts of a city or region; a task that is perhaps not as straightforward as it

62 initially seems. Longitudinal analysis of this nature presents difficulties disentangling the changes associated with the identified activity centres in planning policies from broader measures of urban development. It requires data that can be isolated to small areas with customised extents, is consistent across jurisdictional boundaries, and that has been periodically captured so that changes can be observed over time. Such data constraints limit the suitability of indicators considerably or requires novel approaches to develop suitable data. More positively, the small areas being investigated allow for a degree of precision that is otherwise not possible when investigating entire city areas. Such precision is helpful to resolve common issues with compactness indicators such as accounting for parks and areas of environmental constraints that may distort density calculations (Brewer & Grant, 2015; Churchman, 1999).

Dwelling and population measures The availability of suitable population and housing data is essential to the measurements of the compact city. Census data is an obvious choice as it provides longitudinal data with detailed information on population and housing. This data is methodologically consistent in a census year however geographic boundaries may not be precise in estimating population change over time. This creates problems for accurately studying consolidation in activity centres. For example, the South East Queensland Regional Plan (SEQRP) identifies centres using large circles that are not spatially precise. In their study, BITRE (see description in section 3.2.1) therefore inferred a boundary for each activity centre based on local government plans (BITRE, 2013). Boundaries of this nature will rarely align with census geographical boundaries, so BITRE selected the ABS “destination zones” that intersected with the local plan boundaries. It is not explicitly stated how population and dwelling data was assembled however the report does describe this process as being consistent with the approach used in their analysis of employment, where BITRE (2013, p. 121) states that “population figures were not directly available at the destination zone level, and so were constructed from Census Collection District (CCD) data using an area-weighted concordance”. BITRE (2013, p. 100) acknowledge that when using this approach “…the destination zone containing the centre is significantly larger than the activity centre itself. This means that population estimates for some activity centres may be higher than actual population within the centre.”

These boundary issues are further complicated by ongoing changes to the boundaries of Australia’s statistical geographic standards, including the substantial changes that occurred when the ABS switched from the Australian Standard Geographical Classification (ASGC) to the Australian Statistical Geography Standard (ASGS) with the 2011 census. This creates significant difficulties for researchers examining small geographical scale longitudinal urban

63 change, as comparisons over time on the same geographic area are not immediately possible. Other more consistent geographic boundaries used by the ABS (such as suburb and postal code boundaries) often bare little relation to the on the ground realities of a phenomenon to be studied, or are simply too coarse (Coffee, et al., 2016). The ring based agglomeration method used by Coffee, et al. (2016) provides a simple method to address this problem at lower scales, however their validation criteria could not be satisfied for the very small geographic scales required to understand changes in an activity centre.

One potential solution is the use of a technique commonly known as areal interpolation. Areal interpolation uses the data from source zones (such as census geographic boundaries) to develop an estimate of how that data is distributed in overlapping target zones with different geographic boundaries (Schroeder, 2017). One of the simplest methods of areal interpolation is area weighting, where data is assigned to a target zone based on the proportion of the area that it shares with the source zone/s. The problem with using this approach with population and dwelling count data is that it assumes that population and dwellings are evenly distributed in both the source and target zones (Reibel & Agrawal, 2007). This was the method used by BITRE and it is clearly problematic in urban areas, such as within activity centres, where there can be very abrupt changes to population and dwelling densities. Such abrupt changes with this type of data also reduces the accuracy of the more complex “smoothing” techniques such as those available in geostatistical tools in GIS software (Reibel, 2007).

To overcome these limitations, the area weighted method can be supplemented with additional data such as in the “target density weighting” method (Schroeder, 2007) or the more commonly used method of “dasymetric” interpolation. Dasymetric methods introduce additional auxiliary data sets that assist in more accurately distributing source data in the target zone. The most commonly used form of auxiliary data sources are land cover and land use data, however other sources such as road networks have also been used (Mennis, 2015; Reibel & Agrawal, 2007; Schroeder, 2017). Hawley and Moellering (2005) undertook a comparison of a variety of areal interpolation methods and found that the use of one dimensional road network data proved to be the most accurate. However, this measure is likely to be less accurate in existing areas (such as activity centres) as the pre-established road networks in existing urban areas tend to remain relatively consistent as an area develops. The second most accurate method was dasymetric interpolation using simple land cover data. Hawley and Moellering (2005) detail how the primary shortcoming of this method is related to the auxiliary data not having enough precision to distinguish between high and low density residential areas. The use of more precise auxiliary data in dasymetric methods such as zoning

64 data (Mennis, 2015) or more detailed land cover data identifies categories of residential density (Reibel & Agrawal, 2007), and therefore increases accuracy.

Finding appropriate auxiliary datasets can be difficult in the Australian context. Zoning data in Queensland for example is aspirational; it describes the land use intent for an area and therefore does not necessarily reflect existing uses. As the development assessment system is “performance based”, zoning also does not offer certainty of restriction of particular uses. Freely available land use data from the Queensland Government provides only a single residential category and it has not been possible to obtain current and historical land cover data with pre-existing classifications that distinguish between different residential intensities. It is possible to classify and validate land cover using aerial imagery, however this process can be highly complex (Hussain & Shan, 2016; Reibel & Agrawal, 2007).

Other sources of data can also potentially be used to develop measures for density, particularly data recorded at the address level. In their study of Melbourne, Newton and Glackin (2014, p. 126) make use of the Victorian Government’s yearly published Housing Development Data which provides for “lot-by-lot” data on “dwelling densities, residential yields of development projects, number and location of demolished dwellings, [and] the location of vacant lots”. They found that infill housing development is occurring in suburban Melbourne, however only a relatively small number of infill dwellings (14% of the total number of infill dwellings) can be attributed to policies to cluster development in and near nominated activity centres (Newton & Glackin, 2014, p. 133). The HDD dataset is an excellent source of dwelling information for urban researchers. Unfortunately, not all states freely release similar data at the address level, nor historical versions of the data, and similar data to the HDD was not available in Queensland.

Chhetri, et al. (2013) made use of comparisons between property cadastral data from 2001 and 2006 in their analysis of designated activity centres under the Melbourne 2030 plan. Their analysis found a statistically insignificant correlation between population densities and the plan’s designated activity centres (Chhetri, et al., 2013). As the authors acknowledge however, the use of cadastral data will not detect all new dwellings, particularly certain forms of multiple dwelling structures. This is a severe limitation and unfortunately even the necessary cadastral data is often not available. In Queensland for example, although the current cadastral database is freely available, historical cadastral data is only released for a significant fee.

Other potential data sources include ABS Building Approval data such as used by Buxton and Tieman (2005). Although a good source of data on dwelling changes over time, this data is

65 also based on standard ABS geographic classifications and is therefore subject to the issues previously discussed in relation to interpolating this data to different geographic areas. At the time of writing, this data was available from the SA2 level, with SA1 level data only available for a significant fee. The building approval data is also limited by its coarse grouping of “other residential buildings” which includes all dwellings that are not detached. Buxton and Tieman (2005) note that this data may therefore underestimate rates of medium density housing construction by up to 20%

Planning approvals are another potential source of address level data however, as Woodcock, et al. (2011) highlight, they are often speculative in nature. They therefore do not necessarily identify actual changes to the built form, and the timing of the commencement of construction needs to be correlated with building approvals or direct site observations. It is also possible to estimate dwelling numbers using counts of property addresses data combined with other datasets such as Queensland Government land use data. 2016 address data was obtained from the Queensland Government, however this data does not consistently include all addresses within certain types of unit developments, and includes non-residential addresses such as body corporate addresses and addresses of apartments that are used for short term accommodation purposes.

Employment measures Data that spatially represents employment locations is limited to the ABS’s Journey to Work (JTW) data derived through the census. This data can be obtained in some standard ABS geographic units, but the finest detail is available in a separate geography for JTW data called destination zones (DZN). Unfortunately, DZNs vary in size significantly and include many different land uses including residential uses, which typically make them unsuitable for urban analysis (Day, et al., 2015). Similar to population and dwelling data, spatially precise applications of JTW data therefore require interpolation in order to provide estimates at these smaller geographies (Day et al., 2014). Although methods of dasymetric areal interpolation as described previously are possible, in the case of JTW data this proved to be impractical. This is primarily because of the size of the source geographies which would have required a large number of observations of uses outside of the centre extents. This issue is exacerbated by the poor availability of data from historic censuses. Data from the 1996 census is now no longer released freely by the ABS and instead, customised consultancy service requests must be made. These requests are typically expensive and exceeded the resources available to the project. The 1996 census proved to be difficult to obtain, and the census data eventually obtained from the Australia Data Archive included JTW data at only the statistical local area (SLA) level. These SLAs can be extremely large. The 1996 SLA covering Ipswich for example,

66 covers more than 200km2. The manual creation of an auxiliary dataset for areal interpolation is not feasible at such scales within the resources available for the research. The Queensland government’s land use data is also unsuitable as it lacks richness in terms of the categories of available uses. Proprietary “point of interest” datasets also exist, such as Pitney Bowes’ “business points”, which include points based details about business types and scales (PitneyBowes, 2013). Others have used this data in combination with zoning data to create land use data sets (see Mavoa et al., 2018), however the points are not comprehensive of all business uses, have limited past data available, and are only available on a “fee for service” basis which exceeds the available project resources.

3.4. Conclusion

Theories of plan evaluation suggest that there is value in considering implementation in terms of both outcomes (conformance) and decisions (performance). Absent some form of analysis of how strategic level plans are utilised, it is not possible to discern whether implementation successes or failures are the result of the plan itself, or of its use. Existing empirical studies of the implementation of compact activity centres lack such considerations. They are entirely based on conformance with regional level planning objectives and do not directly consider how these plans perform in terms of their use in the formation of local land use regulations and how these regulations relate to physical development. The previous studies suggest that compact centre policy has proven difficult to implement, with particularly poor results in the nominated centres of car-dependent outer suburbia where such urban forms could potentially yield the greatest benefits. A range of predicted and observed factors have been put forward to explain the lack of implementation of compact city objectives including demographic patterns, employment distribution, property economics, transport accessibility, and the plans and planning system itself. An empirical investigation of the role of plan performance however is still outstanding and research that empirically examines relationships between explanatory factors and the degree of conformance and performance of compact activity centre policy has yet to be undertaken. The existing studies also suffer from a range of limitations. They have typically investigated activity centres as a part of studies on broader trends in urban development and/or only consider changes over relatively short timeframes of less than ten years and have mostly made use of data that is now more than ten years old. There is also a lack of depth in this research with the vast bulk of Australian studies focussed on the case of Melbourne. The only greater Brisbane empirical evaluation of activity centres has suffered from issues in obtaining suitable data that enables the long term analysis of relatively small and discrete areas.

Providing a more up to date look at progress towards compact activity centre policy in greater Brisbane would help to confirm if the previously reported trends of poor conformance are consistent across jurisdictions, and whether these trends have continued in the face of the residential construction boom observed in Australian cities in recent years. New methods can overcome the previously discussed data limitations and can also evaluate the policy over the entirety of its 20 year period. Research evaluating the implementation of the conformance and performance of plans for compact activity centres would better align with theories of plan evaluation, address the specific calls for further research in this field (particularly of compact city policies that seek to alter the urban form), and thereby offer a further contribution to the development of evaluation methods.

4. Research Plan

4.1. Research questions

The previous chapters have shown that in most Australian capital cities, policies to create a polynodal network of compact activity centres are a key regional planning strategy. Compact activity centres are intended to provide all manner of sustainability benefits, from reduced car dependency to improved quality of life. Existing research suggests that after 20 years of policy attempts, there has generally been slow progress towards achieving more compact activity centres. However, there are a number of theoretical and methodological issues associated with these studies resulting in a need for more up to date empirical research that expands to investigate additional cases, and which better considers the use of the planned implementation mechanisms.

This research evaluates the compact activity centre policy that has been in place in greater Brisbane for the past two decades. As described in detail in Chapter 2, this policy seeks to create a network of centres characterised by higher residential densities, a greater diversity of dwelling types, and a mix of employment based uses. The primary implementation mechanism of the policy is to make changes to the land use planning system that enable the desired forms of development. The policy is also intended to be supported through the government provision of infrastructure and services. This process for implementation is illustrated in the left of Figure 8.

Figure 8 – Framework of evaluation of activity centre implementation (author)

The right-hand side of Figure 8 represents the type of evaluation undertaken by this research. The implementation of activity centre policy is firstly evaluated using a conformance based approach to understand if the policy has conformed to actual land use and demographic change in each centre. This provides an updated look at the implementation of centre policy and expands existing research to a new case study in Brisbane. The approach used to undertake this analysis is described in detail in section 4.2.2.

Some argue that strategic level plans are best evaluated on the basis of how they have performed, rather than conformed (Alexander & Faludi, 1989; Faludi, 2000). The research questions therefore also seek to examine how regional level policy for compact activity centres has performed in terms of its use in informing other planning policy. This enables comparisons

70 between the conformance of actual land use changes with measures of plan performance and thereby helps better explain issues associated with plan implementation. Section 4.2.3 provides further details of this investigation and its methods.

Finally, although several authors have put forward a range of explanatory factors for why compact activity centre policy is unlikely to work, there are few studies which compare these factors against outcomes across city wide areas. The previous results are therefore compared to some of these commonly proposed factors to investigate possible relationships between the factors and plan implementation. The methods and approach used to undertake this analysis are described in detail in section 4.2.4.

The following research questions were developed to evaluate the implementation of compact activity centre policy:

• How have greater Brisbane’s activity centres changed in-line with compact city based metropolitan policy?

• How has metropolitan planning policy for compact activity centres influenced land use regulations, and how do actual land use changes conform to these regulations?

• What are the relationships between commonly cited explanatory factors for activity centre conformance with the achievement of compact activity centre objectives?

The following objectives are proposed to address the research questions:

1. Evaluate the conformance of greater Brisbane’s policies for compact activity centres over the past 20 years

2. Evaluate the performance of Brisbane’s regional scale policies for compact activity centres in informing land use regulations, and the conformance of actual land use changes to these regulations.

3. Identify and assess commonly cited factors that influence compact city progress.

4.2. Research Method

This section describes the methods used to address the research questions. The geographic scope of the research and key definitions are firstly provided to spatially identify the location and extent of the activity centres of interest. The methods used to address each research objective are then provided in sequence. The first objective takes a conformance approach of

comparing planned and actual outcomes. The intended outcomes of activity centre policy (section 2.2) are considered in terms of data availability and existing research to identify a series of indicators that are suitable to measure centre compactness at different points in time. These indicators are also considered in terms of their suitability for use in measuring centre intensification, and comparisons to the broader urban area. To overcome the previously discussed data limitations, a new method is advanced that uses Google Street View and aerial imagery to create a detailed land use database. This database is then used in the dasymetric areal interpolation of census data to generate population and dwelling estimates (Limb, et al., 2018). The database is also used to generate employment estimates. The result is the assembly of 15 indicators of centre compactness which are then used to evaluate centre conformance. The second objective addresses plan performance. A method is advanced whereby subordinate local government regulatory land use plans are examined to identify whether their content aligns with regional activity centre policy. This method examines the plans for textual references to activity centre policy, and for changes to regulatory intent. The latter is quantified using “development intensity scores” that provide an indication of how permissive the regulations are for particular types of land use on every property in study area, and whether the regulations have become more or less permissive over time. The reliability of the coding of these scores, as well as the street view observations, were confirmed using a test-retest approach which demonstrated excellent agreement between independently scored random samples. A method to compare the regulatory intent to land use change is also developed in this section. By combining these aspects, it is possible to deduce whether local governments have been making planning decisions in accordance the intended outcomes of regional activity centre policy, and therefore whether there is plan performance. In the final section, indicators to measure the commonly cited explanatory factors for compact city conformance are developed. A method is then proposed to compare these indicators to an overall score of centre intensification using statistical correlations. The results from the application of these methods are then detailed in Chapters 5, 6 and 7, with further discussion in Chapter 8.

4.2.1. Scope of research and key definitions This section outlines and justifies the scope of the research and key definitions. The research evaluates the implementation of compact activity centres in the inner, middle and outer areas within 35km from the Brisbane CBD. The centres themselves are selected based on their inclusion in regional planning policies over time, and their extents are based on a 1,200m walkable catchment from a central public transport node/s.

Geographical extent of research The research evaluates the implementation of compact activity centre policy in the greater Brisbane conurbation. Brisbane is the capital city of the state of Queensland in Australia. It is Australia’s third largest city and has had metropolitan scale policy for compact activity centres in place for two decades (discussed in further detail in section 2.2). Brisbane is therefore considered to be a suitable site in which to study the implementation of compact activity centre policy and provides a “common case” for evaluation. Brisbane has received substantially less attention from researchers investigating the application of the compact city policies in Australia compared to the two next largest cities of Melbourne and Sydney.

Inner, middle and outer locations The research focuses primarily on middle and outer suburban locations, although a single “inner” centre is included in the analysis in order to investigate all non-CBD activity centres in the greater Brisbane area. Middle and outer locations include designated centres in a range of 5km to 35km from the Brisbane CBD. Definitions of what constitutes “middle” and “outer” are subjective and will inevitably change based on the circumstances of different cities. In Brisbane, the 35km range roughly corresponds to the distance the greater suburban conurbation spreads North, South and Westwards from the CBD before changing into distinctly different cities, such as the Gold Coast. The 5km inner/middle boundary has been drawn from the latest version of the regional plan which identifies “the inner 5km” as a distinct area of concentrated economic activity associated with its proximity to the Brisbane CBD (The State of Queensland, 2017c). The boundary between middle and outer for this research is defined as a 15km range from the Brisbane CBD. This range has been selected based on its approximate alignment with the boundaries of the Brisbane City Council local government area.

The research focuses on middle and outer centres due to their prominence in metropolitan policy, the suggested general lack of progress towards achieving their policy intents, and the increased sustainability effects were the policy to be achieved as discussed in section 2.1. The primacy of CBD areas in Australian cities is clear; one need only look at the current crane dominated skylines to see that well serviced inner city areas are becoming more compact with construction of numerous high rise apartment buildings. Such development further reinforces the primacy of the CBD and does not further objectives to create a polycentric urban form. With a polycentric model it is in middle and outer suburbs, where car dependency can be extreme, that viable compact activity centres offer the greatest promise for improving sustainability. The research is therefore focussing on these locations.

Definitions Activity centre For the purposes of this research, the definition of “activity centre” has been drawn from metropolitan policy itself and is defined as an existing or proposed location that metropolitan level planning has identified to “support a concentration of activity including higher density living, business, employment, research, education and services” (The State of Queensland, 2005c, p. 133). As reviewed in section 2.1.1, most Australian metropolitan policy seeks to reshape the existing urban form (characterised by a dominant CBD) to be more polycentric in nature. The activity centres of interest are therefore those that are intended to be of a scale that have significance to the city at the regional level. In the Brisbane conurbation, these centres are spatially identified in metropolitan policy and classified as either “principal” or “major” regional activity centres. The latest iteration of this policy shown in Table 3 provides the following description of these types of centre:

Table 3 - Description of regional activity centre types (The State of Queensland, 2016b)

Centre Type Description Desired population density

Principal regional Principal regional activity centres are key focal points for Within 400m of the centre: 150– activity centre regional employment and critical regional services. These 400 dwellings per hectare centres provide a secondary administrative function to the capital city, accommodating government offices and service Within 400m to 800m of the centres of regional significance. To compete in a competitive centre: global market, these centres support economically significant 100–175 dwellings per hectare clusters that specialise in outward-focused sectors and services, such as professional, health, education, cultural and recreational services. They serve as both creative and knowledge hubs while giving their workforce and resident catchments access to high-order retail and hospitality functions, and cultural and entertainment facilities. As major trip generators, these centres usually have existing or planned, dedicated public transport, such as rail, bus or light rail, and comprise key nodes in the regional public transport system.

Major regional These centres are focal points for subregional employment Within 400m of the centre: activity centre and the delivery of sub-regional services. They 80-200 dwellings per hectare accommodate government branch offices or service centres of subregional significance. They also contain major Within 400m to 800m of the concentrations of business and related activities, cultural and centre: entertainment facilities, and support comparison and 40-100 dwellings per hectare convenience retail uses that meet the needs of their sub- regional catchments. As well as their traditional service roles, growth and commercial development increasingly supports creative and knowledge-intensive businesses to meet the demands of a changing economy. These centres are usually located around key suburban or interurban public transport stations and provide frequent public transport services to link the centre to surrounding communities.

As discussed in section 2.2, although these definitions have changed over time (mostly in terms of the intensity of desired development) the overall policy intent has remained materially similar to these definitions since the inception of regional planning in Queensland in the early 1990s. This is characterised by proposals to encourage higher density residential

74 uses in and near centres, a greater diversity of housing types, and mixed clusters of uses that generate employment and provide localised services. The network of activity centres selected for analysis are therefore those centres within 35km from the Brisbane CBD, that have been designated as Principal and Major Regional Activity Centres since the commencement of the first South East Queensland Regional Plan in 20051. This results in the selection of a total of 21 centres within a 35km radius of the Brisbane CBD (Table 4)2. Most of these centres were also initially identified as a part of the development of the Regional Framework for Growth Management in 1994, however the adopted policy position from 1995 to 2005 only spatially identified the centres of Ipswich and Beenleigh. Almost all the other centres however met the requirements to be “major centres” under these plans (see section 2.2 for more detail).

Table 4 - Selected activity centres for evaluation

Location Name SEQRP Designation Inner Toowong Major Regional Activity Centre Carindale Principal Regional Activity Centre Chermside Principal Regional Activity Centre Indooroopilly Principal Regional Activity Centre Middle Mitchelton Major Regional Activity Centre Toombul Major Regional Activity Centre Upper Mount Gravatt Principal Regional Activity Centre Wynnum Central Major Regional Activity Centre Beenleigh Principal Regional Activity Centre Browns Plains Major Regional Activity Centre Capalaba Principal Regional Activity Centre Cleveland Principal Regional Activity Centre Goodna Major Regional Activity Centre Ipswich Principal Regional Activity Centre Outer Logan Central Major Regional Activity Centre Logan Hyperdome Major Regional Activity Centre North Lakes Major Regional Activity Centre Redcliffe Major Regional Activity Centre Springfield Principal Regional Activity Centre Springwood Principal Regional Activity Centre Strathpine Major Regional Activity Centre

This definition of activity centres aligns with concepts associated with “smart growth” and “new urbanism” which seeks to improve “centering” through transit oriented development principles of increasing mixed-income housing, shops and offices around public transport, as well increased densities around employment rich “edge cities” (Ewing et al., 2002). Research that evaluates centralities in the city form itself examines factors such as population and

1 The SEQRP also includes Specialised Centres and Rural Centres. These centres have more specialised functions and are not necessarily intended to adhere to “compact city” principles. They were therefore excluded from analysis. 2 The greenfield centre of Ripley was still in its incipient stages of development in 2016, and was therefore excluded from analysis.

75 employment density (Ewing & Hamidi, 2014; Giuliano & Small, 1991) and transportation trip densities (Gordon & Richardson, 1996). As this research seeks to evaluate metropolitan policy, it is necessary to define “activity centres” in terms of its usage in the said policy rather than in terms of indicators of progress or the physical variations that may exist in the broader urban form. That is not to discount the impact that such physical aspects may have on activity centre development; it is known that the physical form of activity centres identified by metropolitan policy varies significantly and that this should be accounted for. In Melbourne for example, Yamashita et al. (2006) identified the presence of a typology of regional activity centres that differed from their regional planning characterisations. The South East Queensland Regional Plan only makes the binary distinction of “principal” or “major” centre between what are in fact locations with a far wider variety of land use characteristics. Failing to recognise these differences can be troublesome. For example, BITRE (2013) identified that the most significant recorded progress towards compact city objectives has been occurring in the greenfield centres of Springfield and North Lakes. However, this is more likely to be the result of proportional changes being derived from the centres’ low starting base as greenfield development sites, rather than truly achieving the ideals of the compact city (Goodman & Coote, 2007). It is therefore important that the results of the research are considered alongside the context of the characteristics of the individual centres themselves.

Activity centre extent The spatial extent of activity centres also needs to be defined. Regional plans for Brisbane identify centres using large points that are not spatially precise. Extents shown in local government plans for centres are also not suitable to identify the spatial extent of the centres as they have not been developed with a consistent methodology, and often include large areas of low density residential uses that can be significant distances from the centre’s centre. As the regional planning intent for activity centres (as is also common with compact city principles) is based around transport orientated development principles, the primary public transport interchange of each activity centre was designated as its central point. The concept of walkability to transport and services is one of the core principles associated with achieving the sustainability benefits supposedly inherent with compact urban forms (section 2.1), and has been described in Brisbane’s regional planning policy since its inception (section 2.2). A walkable “neighbourhood” was therefore generated around this point to identify the spatial extent of each centre. Similar reasoning has been used by other researchers to determine influence areas of activity centres (Day, et al., 2015, p. 6).

A distance of 800m is commonly used by planning authorities to determine a “walkable” distance; a distance also used in the latest regional plan to describe the range in which higher

76 residential densities ought to be provided (The State of Queensland, 2016b, p. 36). Canepa (2007) notes how such arbitrary distinctions around walkable distances are overly simplistic and fail to capture the range of issues that also impact on walkability such as surrounding land uses and number of crossings. He explains the relationship between environmental factors and walkability is complex and situational and there is not a one size fits all solution, however he does show some support for considering a larger catchment. This is supported by research from Australia which indicates a significant number of people are willing to access transport beyond the typical 800m range (Ker & Ginn, 2003). They also note the complexity of the environmental situations that contribute to walkability but do not offer a specific alternative range. However, as this research is evaluating plan progress, the range for the purposes of this study needs to reflect metropolitan policy. The standard 800m range has therefore been selected. In their analysis of Melbourne’s activity centres, Newton and Glackin (2014) were able to use the Victorian Government’s centre extents to locate the centres. However, they also examined a number of scenarios that added 400m and 200m buffers to these extents to account for a “ripple effect exerting extra market pressures beyond the boundary onto surrounding properties” from the increased activity occurring in the centre (Newton & Glackin, 2014, pp. 131-132). In this instance, such an approach would also assist in accounting for the higher proportion of residents who are prepared to exceed the standard 800m walking distance to transit as reported by Ker and Ginn (2003). Using the street network to determine a walkable centre extent also has the advantage of excluding areas that are segregated from the central public transport nodes by natural barriers, such as rivers and creeks, and physical barriers such as motorways and large holdings of private property.

The spatial extent of each centre has therefore been defined as an 800m walking distance from the principle public transport facility in each centre, with an additional 400m buffer to capture the centre “ripple effect”. The rationale explaining the selection of the centre transit points is described in section 10.1. The centre transit points were imported into a freely available “neighbourhood generator” tool that calculates walkable distances from points using street network data (AURIN 2016), and polygons of 800m and 1200m walkable catchments for each centre were generated. The neighbourhood generator by AURIN (2016) uses street network data made available by PSMA. As this data set does not account for all pedestrian paths, pedestrian movement for each centre was investigated using Google Street View and Open Street Map. Where additional paths were found, the 800m and 1200m polygons were manually adjusted to account for these movements. A description of these manual adjustments is also provided in section 10.1.

The 2016 version of the Queensland Digital Cadastre Database was downloaded from the Queensland Spatial Catalogue and polygons representing transport infrastructure and water bodies were removed. The adjusted 800m and 1200m walkable catchment polygons were then used to identify the intersecting lots from the cadastre. The selected lots formed the extent of each activity centre.

4.2.2. Objective 1: Evaluate the conformance of greater Brisbane’s policies for compact activity centres over the past 20 years This research objective addresses the research question of how Brisbane’s activity centres have changed in-line with compact city based metropolitan policy over the past two decades. As a conformance based plan evaluation, the methodology takes a positivist position that it is reasonable to evaluate the effectiveness of the implementation of compact activity centre policies by observing changes to the built environment and population, when consistent policies have been in place for extended periods. As demonstrated in section 2.2, this is the case for greater Brisbane activity centre policy, which has now been in place for more than two decades. This research brings together primary and secondary data sources to spatially analyse changes in population, land use, and built form over time to identify the centres that exhibit signs (or not) of becoming more compact. Observations from Google Street View and historical aerial imagery are combined with census data and other secondary data sources to solve issues of data availability that typically inhibit longitudinal analysis of small urban geographies. A range of measures indicative of compact activity centre objectives are then developed based on their suitability and alignment with available data (see Table 9 for a full list of indictors). The centres are then ranked against these indicators to determine their degree of conformance with metropolitan policy. The results of the analysis are described in Chapter 5.

The following sections review common compactness indicators associated with measuring the key characteristics of Brisbane’s compact activity centre policy, namely:

• Higher residential densities;

• Diversity of dwelling types;

• Employment; and

• Mixture of uses and accessibility

The review considers the indicators in relation to their suitability for measuring compact activity centre policy in the Brisbane context, and to the availability of suitable data. Based on this review, a range of indicators are then selected for each characteristic.

Higher residential densities Increased residential densities are a core objective of compact city policies. Measuring density requires population and dwelling figures for the selected areas. Determining accurate population and dwelling figures for the activity centre extent areas however is not a simple a task. In this section, the previously discussed limitations of conventional data sources (section 3.3.2) are addressed using Google Street View data and aerial imagery. A new method is then described that makes use of these sources and dasymetric areal interpolation in order to estimate population and dwelling numbers for each activity centre at different points in time. Common indicators for residential density are then considered and a selection of suitable indicators are selected. These indicators are net population and dwelling density, the average land area of low density dwellings, and the proportion of population living at low densities.

Using Google Street View and aerial imagery to create a land use database Given the extant problems with conventional approaches to measuring population and dwelling density, the research uses observations taken from Google Street View and aerial imagery to create an auxiliary data set that permits areal interpolation of census results. An overview of the key steps involve in this method and further detail on each step is provided in Figure 9.

Figure 9 - Overview of the method used to create a land use database

Record observations and estimates of changes to Identify key nodes and Intersect centre extent land use and built form plot centre extent with cadastral boundaries for each lot (GSV and aerial imagery)

Use census dwelling Observe land use /built structure tables and land Error checking and form changes on lots use database to estimate finalisation of database within all applicable population and dwellings census boundaries numbers

As activity centres are generally small geographic areas, it is possible to manually investigate changes in land use and built form by comparing Google Street View (GSV) images taken at different times. GSV allows for the direct comparison of images from approximately 2007 to

79 more recent images (typically dated from 2015 to 2017). This direct observation of the built environment enables the recording of changes to the built form and land uses. The validity of the use of GSV to undertake built form observations has been demonstrated by a number of previous studies in other areas (Griew et al., 2013; Hwang & Sampson, 2014; Kelly et al., 2013; Rundle et al., 2011; Vanwolleghem et al., 2014). Directly observing changes in the built form between two differently dated images obviously gives clear evidence of change, however in many instances the difference in time between available images is only five to ten years. One of the key benefits of direct observation is that the approximate age of each structure can also be estimated thereby providing historical estimates of the built form that extend beyond the oldest available GSV image. Initial estimates can be further verified through a comparison of historic to current aerial imagery. Geographic Information Systems (GIS) assist greatly in recording and analysing this spatial data.

Using the activity centre extents, new fields were added to the cadastral base data to record the observed GSV and aerial imagery data. Aspects captured included dates of the oldest and newest available GSV images3, estimated building age, whether the land use had changed, current and previous use, current and previous number of storeys, current and previous frontage types, shopping centre names, and whether new development was visible on aerial imagery that was not visible on the latest GSV image.

3 This is necessary to account for possible “spatio-temporal instability of imagery dates” (Curtis et al., 2013)

Of these fields, the estimated building age is crucial in determining estimated land use at different time periods. Figure 10 shows a sample image from GSV. The panel in the top left provides the date of the latest and past imagery and allows the selection of other imagery that was captured between these dates.

Figure 10- Example GSV image (Google Street View)

Where a change to a building or use between the oldest and latest GSV images is observed, a range of years was entered that corresponded to the oldest observed date where the old building/use was visible, and the first date in which the new building/use was visible. In the example shown in Figure 10, the building on the right is under construction in November 2009, but complete in the image available for November 2013. The estimated building age was therefore recorded as 2009 to 2013. Where changes were not visible between GSV image date ranges, historical aerial imagery on Google Earth and historical aerial photographs of each centre (The State of Queensland, 2018), were compared with GSV and the latest aerial imagery (NearMap). Buildings that were not visible on the historical imagery, but which were visible on the oldest GSV image were dated between the year of the historic imagery, and the oldest GSV date. This approach detects building changes but cannot detect whether uses within the buildings have changed between the completion of construction and the first available GSV image. The approach therefore assumes that a building used for a certain category of use is likely to have maintained that use since its construction. Occasionally, GSV imagery would be several years (sometimes up to four years) old. NearMap aerial imagery allowed for changes between the latest GSV image date and the NearMap image to be detected (i.e. buildings that are newer than what is shown in GSV). In these instances the use was estimated on the basis of the buildings’ appearance from the aerial, as well as from the

81 oblique angle aerial imagery available on NearMap which can provide a view of a building’s elevation (the 3D building layer on Google Earth offers a free alternative to the subscription based NearMap for this function). Buildings that were in the oldest GSV image and the oldest historical aerial image were recorded as “Pre 96”4.

Once all lots have been observed, the result is a database that identifies land use change to the address level, over a 20 year period. Analysis of this data alone can provide details on a range of factors relevant to progress towards compact activity centres, such as ratios of residential to non-residential land, and changes in areas of different types of residential and non- residential uses. It can also be used as auxiliary data in dasymetric interpolation of census data in order to estimate the population and number of dwellings, and to calculate a range of density types.

Estimating population and dwelling numbers To interpolate population and dwelling estimates from census data, it was firstly necessary to develop estimates for land use changes in a broader area. The CD or SA1 areas for each census year were selected where lots within the activity centre extent were located within these areas. The maximum extent of the combination of the CD and SA1 areas for each census year was then used to select lots from the cadastre layer. NearMap and historical aerial imagery were then used to assign an estimate of the land uses for each of these lots, and to also record a date range of when obvious changes to the built form occurred on these lots. These estimates were then combined with the activity centre database, and the building age fields and imagery dates were used to create snapshots of the estimated land use at the time of the 1996, 2001, 2006, 2011 and 2016 censuses. As these dates were typically recorded as a range of years, the median year of the range was used to assign the change of use to one of five cohorts that align with census years. The combined table was then spatially joined to the applicable census boundaries for each census year, and the area of land for the various types of residential uses were summed for each census boundary. The census dwelling structure definitions were then matched to their corresponding categories in the GSV database (see Table 6, p88). The proportion of area of low (detached dwellings and duplexes), low-medium (townhouses, caravan parks and apartments up to two storeys), medium (apartments up to three stories), and high (apartments of four or more stories) density residential land within the activity centre extent vs. outside the activity centre was then calculated for each census boundary and year. Census dwelling structure data for counts of private occupied dwellings and population that coincided with these types of residential uses were then multiplied by the relevant proportional

4 This research is focussed on a 20 year period however this method could also be used to estimate older building dates on the basis of style and/or through the use of older aerial imagery

82 figure to determine an estimate of the number of dwellings and population within the activity centre extent in each census boundary area. In the 1996 census, unoccupied dwellings are estimated using the same method, however census counts of unoccupied dwellings in the 2016 census are only provided as a total figure. Unoccupied dwellings were therefore estimated using the proportion of overall residential land in/out of the activity centre. The results were then summed within each centre to develop an estimate of the number of dwellings and population within the activity centre extent at the desired census years (in this case, 2016 and 1996).

Limitations and reliability The method described here has several limitations that ought to be considered. Lot by lot manual observation is tedious and time consuming, thereby limiting the extent of areas that can be observed. As already mentioned, where uses cannot be directly observed, the approach requires more subjective estimations of use that rely on assumptions based on surrounding uses, and that a building has maintained a similar type of use category since its construction. Fewer estimations will be necessary as more GSV data becomes available with time, provided GSV data continues to be made available on the same terms. Bader et al. (2017) discusses this matter of proprietary data further, as well as other matters related to ethical concerns and “rater fatigue”. As a manual process, there is of course the possibility of missing changes, misinterpreting land use and/or building age, and introducing errors through the data entry process. Checking the data for logical consistency allows for the identification and correction of many of such errors, however some are inevitable.

New computer processing methods hold promise for overcoming some of the limitations associated with manual observation and data entry. Naik et al. (2017) have utilised automated processing of GSV historical images to reliably record changes in urban appearance over time. Though technically complex, the use of such automated methods to determine changes in land use and built form would enable the comparison and analysis of greater numbers of activity centres and geographies. Initial attempts at the automated classification of detailed land uses by Li et al. (2017) show some progress, however their testing revealed issues with accuracy when distinguishing between different types of residential structures. Until such issues are resolved however, using observations of GSV and aerial imagery as described here enables the detailed longitudinal analysis of discrete areas, such as activity centres, and also assists with the areal interpolation of other data sources such as census data.

Agreement tests were performed to assess the reliability of the data observations that make up the GSV land use database. For nominal data types, such as that used in the GSV land use database, Cohen’s Kappa (κ) is a commonly used coefficient that compares the results of two

83 observations of the same object (Cohen, 1960; von Eye & Mun, 2005). κ records the extent that two observers agree or disagree beyond chance on a scale of 1 to -1 where 0 indicates an agreement no better than chance. The standard conditions required to measure κ require the units to be independent, the categories of measurement to be independent, mutually exclusive and exhaustive, and that the observations are made by independent judges (Cohen, 1960). These conditions can all be met for an assessment of the GSV database. As project resources did not permit the recruitment of an independent coder, the condition of judge independence was met using a test-retest approach, with a lag time of more than a year between testing. Independence in test-retest approaches is considered to be valid where a sufficient time period has elapsed between tests, and the status of the observed object has not changed between observations (Allen & Yen, 2001; Bujang & Baharum, 2017). These conditions are met by the extended period between tests, and the fact that the observations were made from images taken at a specific time which therefore cannot have changed. Project resource limitations also prevented the re-observation of the entire database, so a random sample of 50 records were selected for the re-test based on standard sample size formulas for κ (Cantor, 1996; Gamer et al., 2019). The initial observations were compared with the re-tested sample observations for agreement in five key aspects from the GSV land use database: 2016 land use type, 1996 land use type, cohort of land use change, building height, and building frontage type. The results of the agreement tests are shown in Table 5 below which include κ and its confidence interval, the coefficient of raw agreement, and κn (Brennan & Prediger, 2016; von Eye & Mun, 2005).

Table 5 - Results of reliability testing of GSV land use database observations

Observation type Cohen’s κ Raw agreement ra Brennan and Prediger’s κn 2016 land use type 0.884 (95% CI, 0.780 to 0.987) 0.920 0.910 1996 land use type 0.905 (95% CI, 0.805 to 1) 0.940 0.932 Cohort of land use change 0.794 (95% CI, 0.623 to 0.964) 0.920 0.90 Building height 0.881 (95% CI, 0.651 to 1) 0.980 0.970 Building frontage type 0.810 (95% CI, 0.612 to 1) 0.960 0.946

Using the categories for κ identified by Fleiss (1981), all types of observation demonstrated excellent agreement between tests and were statistically significant beyond chance agreement (p < 0.05).

The areal interpolation method of incorporating census results also has its limitations. The results of interpolation are estimates only, and although accuracy ought to be significantly increased through the dasymetric approach of aligning auxiliary land use data with census dwelling structure data, there will be errors introduced through this process. As the method relies on weightings of areas, this is most likely to occur when significantly larger or smaller lots (compared to the typical lots within a given source area) are unequally located within or

84 without the extent of the target area. The use of census data itself also has some potential issues. The ABS (2016) uses “introduced random error” to protect the privacy of census respondents. These errors are proportionally greater on data consisting of smaller counts. As population and dwelling figures are derived from summing specific fields of the dwelling structure data tables, some of which have small values, the summed census data is therefore likely to be subject to some degree of random error. However, as there are relatively few census areas covering each of the centres with very small numbers of population and dwellings, random errors are likely to be small in relation to the summed total results for each activity centre. The dwelling structure data also only counts “private occupied dwellings”, therefore excluding counts of residents in short term accommodation, hospices and other “non-private” dwellings (ABS, 2016). Issues have also been identified where observed dwelling types do not align with the census counts for that type of dwelling; a situation that has been explained by discrepancies in how ABS field officers interpret dwelling structure types (ABS, 2017).

Measures of residential density Existing studies make use of one, or a number of expressions of residential density. The most coarse level of density are indicators for gross population and dwelling density are expressed as the number of persons or dwellings divided by the total area of the studied geography (Burton, 2002; Ewing & Hamidi, 2014; Song & Knaap, 2004; Stathakis & Tsilimigkas, 2014; Yang, 2008). This measure is relatively simple as it can use the area of the component census geographies. More precise, are measures of density expressed as the number of persons or dwellings divided by the total built up area of the studied geography (Abdullahi, et al., 2014; Burton, 2002; Ewing & Hamidi, 2014; Lin & Yang, 2006; Stathakis & Tsilimigkas, 2014). Definitions of what constitutes “built-up” vary, but they typically exclude undeveloped sites, parks, and other “non-urban” uses. This measure requires additional data sets in order to identify the built up locations. The most precise measure of typical density indicators is “net density”, which measures the number of persons or dwellings per area of residential land, and includes only the lot areas of land used for residential uses (Churchman, 1999). This measure was used by Abdullahi, et al. (2014); Burton (2002); and Stathakis and Tsilimigkas (2014), and also requires additional datasets in the form of locations of residential land uses, and lot boundaries.

The development of the land use database and estimates of population and dwellings described previously, enables the measurement of any of these density indicators. The definitions of density in the regional plans have changed slightly over time, from the net density definition in the RFGM plans, to a revised definition in the SEQRP plans that is more akin to “street density” (see Churchman (1999, p. 391)), where the area is to include local

85 roads and open space. This later definition is imprecise as it is difficult to tell which parks or roads are to be included. The measure of net density has therefore been selected as it is generally aligned with the regional policy definitions and is the most precise measure of actual residential density. Although others use multiple measures of density (e.g. (Burton, 2002), the other measures have been excluded as they will preference areas with larger proportions of residential land within the centre bounds.

Other density indicators relate to measures based on low density housing forms. Song and Knaap (2004) use indicators of the median floor area and median lot area of “single family dwelling units” as a part of their measures of density. Data is not available for residential floor area so this measure could not be included. Measures of the lot area of single houses provide a good indication of the relative intensity of low density residential uses. The nature of the data available through the land use database, means that a median cannot be calculated. However, as the amount of area of low density uses is known from the database, it is possible to calculate the average amount of land per low density dwelling. This form of density is often used in Queensland land use regulatory provisions for subdivisions. Ewing and Hamidi (2014) include a measure of the percentage of the population living at low residential densities, which is relatively simple to calculate using the land use database or interpolated census data.

Lin and Yang (2006) and Abdullahi, et al. (2014) use building density as a measure of compactness. This measure divides the building footprint by the built up land area. This measure was investigated, however in Brisbane’s activity centres it showed little relationship with other measures of compactness and instead favoured centres containing a greater number of uses with large building formats. Used on residential uses, it captures the building footprint only and therefore does not measure residential intensity, such as high rise residential uses. Measures of residential plot ratio (i.e. the amount of residential floor area divided by land area) would be more useful however, as previously described, residential floor area data was not available.

Based on data availability and suitability to measure the objectives for Brisbane’s compact activity centre policy, the following residential density indicators were therefore selected for use:

• Net population density (persons per net residential hectare)

• Net dwelling density (dwellings per net residential hectare)

• Average land area of low density dwellings (Total net land area of low density residential uses divided by the number of low density dwellings)

• Proportion of population living at low densities (population of low density uses divided by total centre population)

Changes to each of these measures over time can be calculated in both absolute and relative terms. As the geographic extent of each centre is constant, the relative change in population and dwellings can also be considered as indicators of residential intensification.

Diversity of dwelling types The review of compact activity centre policy in section 2.2 showed that increasing the diversity of types of residential uses in activity centres has been a recurrent theme in metropolitan policy since its inception. This is typically considered in terms of desired changes from low density residential uses such as detached dwellings, to “higher” density uses such as townhouses and apartments. What constitutes “higher” density has intensified as regional policy has evolved over time, from the “medium density” uses envisioned in the RFGM plans (33 dwellings per net hectare), to 120 dwellings per hectare and more under subsequent plans. Reductions in low density uses and increases to medium and high density uses are therefore considering in keeping with regional planning policy. The regional plans also seek a variety of housing types “…to match the changing needs of the community and changing household size and structure” (The State of Queensland, 2005c, p. 65). In this section, indicators for changes in dwelling type are reviewed and suitable indicators are selected. The selected indicators measure the proportion of different dwellings types, and the Index of Qualitative Variation (IQV) of these proportions.

Dwelling structure data is required in order to measure these aspects. As discussed in the previous section, census data includes dwelling structure data however there are difficulties using this data for customised extents (such as the walkable catchments this research uses for activity centres), and also with changes to the geographic boundaries of census data over time. The use of the land use database derived from Google Street View observations as an auxiliary dataset in dasymetric areal interpolation allows for the estimation of the number of dwellings and population associated with the various dwellings types observed. When creating the land use database from GSV observations, residential land uses were categorised to allow for alignment with standard census dwelling structure types. Table 6 shows the composition of dwelling categories and their relationship to the use observations made using GSV and census dwelling structure data.

Table 6 - Dwelling type classifications

Dwelling type GSV Use Type GSV Use Definition Census dwelling structure type/s

Low density Detached Dwelling Single dwelling (or dwelling with granny Total of Detached dwellings, residential flat) on a single lot apartments attached to house dwellings, dwellings attached to Duplex Two dwellings on a single lot a shop/office and not stated dwellings

Low-medium Multi Unit More than two dwellings on a site, Semi terrace/townhouse density Development Low predominantly low rise, and typically dwellings, one or two storey residential townhouse style units, or 2 storey apartments, apartments Caravan/cabin/houseboat dwellings, improvised dwellings

Medium density Multi Unit More than two dwellings on a site, up to Three storey apartments residential Development Mid three storeys, and typically apartment style units

High density Multi Unit More than two dwellings on a site, over Four storey or more apartments residential Development High three storeys, and apartments Mixed Use Apartments with ground floor Residential commercial uses such as offices, retail and food and drink.

Residential Complex Very large site with a range of residential densities from high rise apartments, to townhouse development. Staged development, constructed over several years. Mixed Use Complex A mix of office and residential uses (including short term accommodation) where there is a similar amount of residential and office uses. Also includes additional ground floor commercial uses such as office, retail, food and drink etc.

Studies of urban compactness typically do not include dwelling mix categorisations however several studies do include indicators (typically categorised as density or mixed use indicators) that are suitable for dwelling mix. These are generally expressions of different dwelling structures as a proportion of all dwellings such as the proportion of higher density dwellings compared to the proportion of lower density dwellings (Burton, 2002; Stathakis & Tsilimigkas, 2014; Yang, 2008). Yang (2008) also makes a calculation of the Index of Qualitative Variation (IQV) between the proportions of different housing types. IQV is calculated using the following formula where K = the number of categories, and Pct = the proportion of each category expressed as a percentage.

5 source: (Yang, 2008)

This formula results in a score of 1 where the proportions of the different dwelling types are perfectly mixed and approaches 0 when less well mixed. For example, using the four dwelling type categories defined in Table 6, the IQV for a centre would equal 1 when the proportion of dwellings in each category is 0.25. The IQV would equal 0 if a single category had a proportion of 1 and the other categories had a proportion of 0. This measure is useful in understanding overall variability in different dwelling types, however it does assume a normative case that having a perfect mix of the selected categories is desirable.

These indicators are all suitable to measure the objectives for dwelling mix in compact activity centre policy, and data can be derived to use them. The following measures have therefore been selected:

• Proportion of low density dwellings (the lower the proportion, the more compact the overall dwelling types)

• Proportions of low-medium density dwellings, medium density dwellings, and high density dwellings (the higher the proportion, the more compact the dwelling types)

• IQV of the above categories of dwelling types (the closer to 1, the more mixed the dwelling types)

The absolute and relative change of these ratios can be considered over time and compared to non-centre areas to give an indication of how centre dwelling mix is changing compared to changes to the broader urban area. However, creating measures of change to the proportions is somewhat problematic. This is primarily due to how changes in one area effect another. For example, a centre that shows some growth in medium density development and higher growth in high density development may actually record a negative score for the proportion of medium density development. Similarly, measuring changes to the IQV of housing types must also be considered with caution. If increases to the IQV are considered as being an improvement, centres that were already well mixed at the earlier point in time and which maintained their mix, would not show significant improvement in the IQV score as it would have already been close to the maximum of 1.

Taking the relative growth of each category of dwelling type as a proportion of the total existing dwellings in 1996 is used instead. The assumption behind this measure is that centres that show overall relative reductions in low density dwellings, and increases across a range of other dwelling types, will be best exhibiting the type of mixed dwelling growth expected of more compact centres.

Employment Policies to improve employment outcomes in centres are perhaps the most enduring component of activity centre policies, existing before conceptions of the compact city were developed. As discussed in section 2.1.1, these policies were considered in terms of providing work to nearby residents, as well as in terms of the provision of a diverse array of localised services. Initially justified in terms of economic and infrastructure efficiencies, the subsequent inclusion of compact city based objectives saw justifications evolve to include sustainability factors such as reductions in vehicle trips/distances. Generating employment continues to be a key aspect of Brisbane’s activity centre policy. This section details the challenges involved in obtaining suitable employment data, and how combining observations using Google Street View, aerial imagery, and pre-established workspace ratios can overcome these issues. It also examines existing measures used to evaluate employment in terms of the compact city. Based on data availability and the suitability of measures, indicators of net job density, average employment intensity, and employment plot ratio were selected to evaluate centre conformance in this category.

Employment data The lack of suitable data (section 3.3.2) drove the pursuit of a different approach to estimating centre employment. By using Google Street View observations, it is possible to make a range of other observations of the built form such as building heights. By combining observations of land use, height, and building footprint areas it is possible to develop estimates of the floor areas of non-residential uses. These floor areas can then be combined with existing data of typical employment by floor area in order to create an estimate of the employment capacity of the built form.

Estimating floor area from building footprints Building heights were observed using GSV (both current and old) and recorded against each building. As this height was to be used for calculating floor areas, the number of observable stories being used for the stated use was recorded (e.g. car parking stories were to be excluded). Developments with buildings of multiple heights were given the average height of the use. For example, a shopping centre where approximately half the footprint is 2 storeys, and the other half is 1 storey, would have a 1.5 height. In more complex situations, the use was split into different polygons to record the footprint areas of the different levels. Older hotels were given a single story for GFA purposes as they typically consist of a single storey for the primary use, with very limited short term accommodation on the upper floors. The other exceptions were industrial mezzanines, where industrial buildings have a two storey front office, which often

90 extends as a mezzanine over the industry use. These types of industrial buildings were recorded as 1.5 storeys.

The 1996 building footprints were traced from historical aerial images made freely available through the QImagery website (The State of Queensland, 2018). A total of 53 photographs of the centres were sourced from this website and manually georeferenced in ArcMap. The images were selected based on their scale and dates. Although widely available, images with scales greater than 1:25,000 were not suitable due to their poor resolution. Where available, higher resolution images were selected. It was rare to find images dating exactly to July 1996 so the images that offered an acceptable resolution, at a date closest to this date, were therefore selected. Depending on the coverage of the images, sometimes multiple dates were needed to cover a centre’s extent. This resulted in a mixed range of image dates from November 1994 to November 1997. The historical imagery was used to observe differences in building footprints between 2016 and 1996, in particular:

• The absence of a significant structure visible in 2016;

• Structures that were different to those recorded from the 2016 observations (excluding dwelling houses - for detached dwellings, changes to their size through extensions or even a complete rebuild is not of interest to the research. A new house however represents a new dwelling, and these footprints were therefore recorded.); and

• The presence of structures that were not visible in 2016.

Where differences were found, the older footprint was traced.

The useable floor area of each building was calculated by multiplying the shape area of each footprint, by the number of stories, and then multiplying by 0.75. The 0.75 figure is used to account for circulation areas, toilets, building eaves, entrances, stairways, and other parts of the building not used directly for the use. In some instances, more precise figures were required. For example, mixed use developments typically contain floors of apartments above shop fronts. These shop fronts however do not cover the entire footprint of the building, so using the entire building footprint would therefore overestimate the employment area. In these instances, more accurate floor areas were obtained from development plans and/or reports, third party websites about the development, or if not available through these sources – estimating the footprint of the employment area by GSV and aerial image observations. Shopping centres represent another type of use that required more precise floor area figures, particularly for large format centres with complex internalised arrangements of multiple levels and car parking. The Property Council of Australia publish floor area statistics of shopping

91 centres (Property Council of Australia, 2016). The 2016 version of this data was joined to the building footprints, and the total gross floor area was used in place of the estimated floor area figure. For the 1996 footprints, shopping centre floor areas were obtained from a similar document dated to 1993 (Building Owners and Managers Association Queensland Division, 1993). Where required, the floor areas were updated to account for development that occurred between 1993 and 1996 (see section 10.3 for further details about these adjustments).

In addition to providing approximate floor areas for determining employment potential, the building footprints can also be used to measure a range of other changes to the built environment. Without building footprints, the GSV land use database would be based entirely on cadastral boundaries. This approach can record use changes, but changes in the scale of the use would be based on land area only. Land area changes can potentially be misleading. For example, with a cadastre based approach alone, a small building on a very large site would indicate a more intensive use than a large building on a small site. Building footprints give additional data to help differentiate these aspects of development. Additionally, where the use of a property had intensified but otherwise had not changed, the building footprint areas can reveal a change in intensity and scale that would have otherwise not have been observable.

Although it is possible to automatically classify building footprints from aerial imagery, there were difficulties in obtaining freely available, cloud free images, at suitable dates and resolutions, across all of the centres. Other difficulties with automated classifications relate to footprints being drawn over cadastral boundaries (creating future issues when attempting join cadastral data to the footprints), and the inability to differentiate areas used exclusively for car parking. Some building footprints are freely available through Open Street Maps, however at the time of the research these footprints were not comprehensive. Google maps also has building footprints which can potentially be “scraped”, however these are proprietary and such actions are of questionable legality. Commercially available footprints also exist, however the cost of these services was prohibitive. 2016 building footprints were therefore created by manually tracing aerial imagery from Nearmap images dated as close as possible to July 2016.

Estimating employment By combining the floor area data, with land uses, and employment floor area use by industry data, it is possible to estimate the amount of employment per building. The City of Sydney painstakingly conducts “Floor Space and Employment Surveys”, in which they manually measure all employment based uses within their local government area, and produce reports on employment and employment floor areas (City of Sydney, 2012a). As a part of these surveys, “work space ratios” are calculated which detail the number of employees per floor

92 area based on the industry type. The council publishes 10 separate reports for different parts of the city. Perhaps unsurprisingly, ratios vary considerably between different locations and especially between non-CBD and CBD based areas. The ratios from seven non-CBD based areas (selected due to being more comparable with the uses and built form expected in non- CBD activity centres) were therefore used to calculate the median work space ratios across these locations (City of Sydney, 2012b, 2012c, 2012d, 2012e, 2012f, 2012g, 2012h). The most appropriate industry or space use ratio was then assigned to each land use type in the GSV land use database. As the land use database does not distinguish between individual uses within shopping centres, shopping centre uses were assigned the average ratio of the retail, food & drink, and office ratios to account for the most common use types contained within them. Hospital uses were also provided a different ratio to account for their higher rates of employment. The 2016 JTW data includes a single DZN covering the Prince Charles Hospital at Chermside. As floor area of the hospital has been estimated, it was possible to calculate a hospital employee ratio by dividing the floor area by the number of employees listed in the DZN. This rate (16.8m2 per employee) was then applied to all hospital uses in the land use database. The floor area of each building was then divided by the assigned employment ratio in order to create the estimated number of employees.

The measure’s main limitation is that it cannot measure actual economic activity. For example, depressed areas with vacant or poorly performing uses would yield the same number of employees as economically vibrant locations. It is therefore best to consider this measure as an indicator of the employment capacity provided by a given built form and land use. None the less, the measure does show concordance with the JTW employment numbers across a number of locations. The 2016 census has a total of 11 DZNs that contain all employment uses within the activity centre area extents. The total employment numbers from each DZN can therefore be compared with the employment estimation generated from building floor area estimates. Based on this sample, the floor area estimate tends to slightly underestimate employment (Table 7). Lin’s concordance coefficient is a statistical test that is commonly applied to check the strength of association between two different methods of measurement of the same phenomenon. Applying Lin’s concordance coefficient (Lin, 1989) to these figures shows a substantial to high degree of concordance between the measures (Rc = 0.901 (95% CI, 0.690-0.9716)).

6 Calculated in SPSS using a script created by Garcia-Granero (2005) (https://gjyp.nl/marta/Lin.sps)

Table 7 - Census employment vs. footprint estimated employment

DZN Centre 2016 Census Footprint Estimated Location Employment Employment Carindale 4242 4357 Chermside 2035 1089 Chermside 5306 5314 Chermside 5424 5533 Indooroopilly 5048 4907 Ipswich 3818 4045 Ipswich 5103 4879 Strathpine 3562 3276 Toombul 1097 1346 Toowong 5087 3314 Upper Mount Gravatt 5991 5318

Total 46713 43378

Measures of employment Employment related measures are often used in measuring the degree of mixed land uses (Ewing & Hamidi, 2014; Lin & Yang, 2006). Such measures attempt to gauge proportions of residents or dwellings to jobs or jobs of different sectors. Although useful in the analysis of broader geographies, these types of indicators are less useful when investigating discrete urban areas in which employees are likely to be drawn from residences beyond the area of interest. Measures such as the ratio of employees to residents will naturally preference (assuming a higher ratio is better) centres with higher proportions of residential uses. Relative and absolute changes in numbers of jobs, as well as employment density are also commonly used employment related measures (Day, et al., 2015; Ewing & Hamidi, 2014; Lin & Yang, 2006; Liu et al., 2018). Like the previous discussion on residential density, employment density can also be measured in terms of gross, built-up, or net areas. As the GSV land use database allows for the identification of employment based parcels of land, net employment density (the number of jobs divided the number of hectares of employment related land) is considered to be the most precise measure of employment density. Also like the discussion of measures for residential density, plot ratio is also a useful metric. In the case of employment related uses, the development of estimates of floor areas enable the overall ratio of employment floor area to employment land area, to be determined for each centre. This measure captures the relative built form intensity of employment based uses. Measuring the floor area in relation to the number of jobs is also useful as it gives an indicator that favours centres with more job intensive uses, such as the uses preferred by metropolitan planning policy.

The following indicators have therefore been selected to measure centre employment outcomes:

• Net job density (the number of jobs divided by the area of employment land in hectares)

• Average employment intensity (the floor area of employment buildings divided by the number of jobs)

• Employment plot ratio (the floor area of employment buildings divided by the area of employment land)

Changes to each of these measures over time can be calculated in both absolute and relative terms and as the geographic extent of each centre is constant, the relative change in total employment can also be considered as an indicators of employment intensification.

Mixture of uses and accessibility Creating a greater diversity of uses within activity centres has been a continuous objective of Brisbane’s metropolitan policies for centres, particularly in terms of mixtures of higher density residential, retail, office, community and entertainment uses. Like broader conceptions of the compact city, this approach is justified in terms of sustainability outcomes through the reduction in vehicle trips resultant from multiple-purpose journeys and more localised employment and services, and economic benefits related to “clustering” of multiple businesses and services. Brisbane’s metropolitan policies have also included proposals to improve the streetscape amenity associated with mixed-use development to further encourage walkability. Measures of mixed use are subject to similar constraints surrounding data availability, however much of these have been overcome through the development of an address level land use database created using Google Street View observations and aerial imagery. Typical indicators for measuring mixed-use in urban areas are reviewed and selected based on their suitability for measuring compact activity centre policy within the constraints imposed by the available data. Measures of land use variation, average Euclidean distance, median residential proximity, proportional proximity of residential uses, and a new measure to quantify active frontages, were selected for this purpose.

Mixed use data Measuring mixed-use areas in a meaningful manner requires relatively detailed land use data. This is especially true when examining the small geographies associated with activity centres. The availability of suitable land use data continues to be an issue in Australia, with researchers making use of proxy datasets (such as zoning data) or paid commercial services (Mavoa, et

95 al., 2018). These issues have already been described at length in previous sections. The development of a detailed land use database using Google Street View and aerial photography was also previously discussed. This work provides a record of estimated land use change over time that solves many of the difficulties in obtaining suitable data to longitudinally measure use mixture. Google Street View also permitted observations of type and length of building frontages, which can be used to give an indication of progress towards outcomes for improved streetscapes.

Measures of mixed use Measures of mixed use can typically be considered in terms of indicators that examine matters of centrality, direct land use mix, or proximity and accessibility. Centrality measures include indicators of relative variations in population and employment density, changes in density gradients from CBD locations (Ewing & Hamidi, 2014), and measures of the physical shape of cities such as “coreness” which examine the ratio between a city’s “core” and built up areas (Stathakis & Tsilimigkas, 2014). These measures of centrality are primarily applicable when investigating entire metropolitan areas and are less useful when examining centre areas based on a walkable catchment which, by their very definition, are already centred.

More direct measures of mixed use have a range of scales of analysis. At the most precise, Burton (2002) measures the amount of individual uses that are, literally, mixed. This takes the form of the ratio of retail floor area that is part of a development with residential uses to all retail floor space, and the ratio of the number of mixed-use residences to all other flats. By using a combination of the land use database, estimated floor areas, and address point data, it was possible to use these measures for the activity centres. In 2016 half of the centres however did not have any actual mixed use developments, and in 1996 only two centres had this form of development. Although literal mixed-use development is desired in centres, the metropolitan level policy is aiming for more area wide effects. Other measures proved to be more suitable for measuring this objectives and the measure was therefore abandoned as a measure of policy conformance. Burton (2002) and Mavoa, et al. (2018) used the count and spread of key facilities as a mixed use indicator, but these forms of measure also focus on precise aspects of the built form that do not capture the broader intent of the compact activity centre policy.

As previously discussed, some authors use employment based indicators to identify mixed use areas. As employment is considered a separate category of measurement for compact activity centre policy, suitable employment measures have been included in the employment section. Also previously discussed were the use of ratios such as residential to non-residential land (Burton, 2002; Song & Knaap, 2004; Stathakis & Tsilimigkas, 2014). These forms of ratios

96 can be improved by the inclusion of additional use categories. Song et al. (2013, p. 11) reviewed and tested an extensive number of mixed use measures and concluded that Entropy Index measures are most appropriate where more than two categories of land use are to be considered and “the unit of analysis is an appropriately small scale to detect the influence of land use mixing”. The land use Entropy Index is often used to quantify the degree of mix between uses (Ewing & Hamidi, 2014; Lin & Yang, 2006; Liu, et al., 2018; Mavoa, et al., 2018). Entropy is defined using the following formula where Pj is the percentage of each land use type j, and k is the number of land use types j (Song, et al., 2013).

Like the IQV score discussed previously, as the score approaches 1 the more even the input proportions, and the score approaches 0 as categories become less even. To measure the mix of uses appropriate to activity centre policy, the land uses in the GSV database were assigned categories of residential, retail, office, or community. The land area associated with each use category was summed for each centre. Proportions of each summed use category were then used to calculate the Entropy Index for each centre. There are several disadvantages to the Entropy Index. Firstly, it assumes that a perfect mix is desirable, so the results are highly dependent on the choice of suitable input categories. In this instance these categories generally align with overall metropolitan policy intent, however the policy did not specify the exact intended proportions of different use types. ENT is also “symmetric” and therefore will return the same result with different distributions. For example, a distribution of 50%/5%/15%/30% would give the same result as 5%/15%/30%/50% distribution (Song, et al., 2013). The measure also does not give an indication as to the spatial distribution of the use categories. The justification used for a mix of uses is heavily reliant on the benefits of these uses being proximate to each other. For these reasons, the Entropy Index alone does not suffice as a suitable mixed-use indicator in the context of evaluating compact activity centre policy.

Proximity measures can measure the spatial relationships between different uses. Abdullahi, et al. (2014) used standard Euclidean distance tools in ArcMap to determine average distances between different use types. This method works by creating a grid where the distance from the centre of each pixel is measured to the nearest pixel of interest. Different grids can then be combined into a mosaic grid that combines the distances recorded to pixels of interest. Where pixels of interest represent certain land use types, pixels with lower total distances representing

7 Source: (Song, et al., 2013)

97 areas that are more proximate to the different use types and are therefore considered more “mixed”. This method was applied using a 7x7m grid, with distances being calculated to pixels created from building footprints and classified as residential, commercial (office and retail), and community. In the final mosaic grid, the pixels that fell within the built up area of each centre were selected, and the average pixel distance was determined for each centre. A disadvantage of this method is that it measures direct distance and therefore ignores possible barriers to movement such as private property, roads, rail lines etc.

Song and Knaap (2004) use distances between residential uses and other features. Their measure takes the median distance of all residential uses to the nearest commercial use, bus stop, or public park. However, their study also used Euclidean distance, which therefore fails to account for barriers to movement. By using ArcMap’s network analysis tools, it is possible to use the same method but with distances that reflect the actual road network. Road and track centrelines were obtained from the Queensland Government’s open data portal8, and this data was filtered to remove paths not accessible to pedestrians such as motorways and busways. Although this dataset includes most bikeways and pedestrian paths, it is not comprehensive for these features. Aerial images were used to identify unrecorded paths that would impact on path calculations and the network was manually edited to include these features. As done by Song and Knaap (2004), origin features were defined as the centroid of the building footprint of residential uses. For multiple dwelling residences, address point data was used as origin points. As discussed previously, this data is not perfect (see the previous section discussing Higher residential densities) however it is the best available measure of the number of dwellings on each multiple dwelling site. Destinations were defined as the centroids of commercial uses, community uses, and parks. As the centre areas are defined by their distance from a central public transit node, distances to transport were not included. The path- based closest distances between origins and destinations were then calculated for each origin point, and the median distance of all origin points for each centre was determined. The average of these median distances between commercial, community, and parks were then calculated to determine a single overall proximity indicator. To further measure pedestrian access, Song and Knaap (2004) also measure the proportion of residences within 400m of commercial uses. This indicator is relatively simple to develop by combining the dwelling estimations calculated previously with the data from the origin-destination calculations, and was also included.

8 https://data.qld.gov.au/dataset/baseline-roads-and-tracks-queensland

Measures of streetscape design in relation to evaluation of the compact city is an aspect with few existing studies. Talen and Jeong (2018) consider these aspects in terms of their study on main streets however their analysis is primarily based on factors relating to the presence of selected land use types. For this study, the previously discussed mixed use indicators are considered to already address matters relating to appropriate land uses. To capture the design aspects related to main streets, Talen and Jeong (2018) also include “quality” indicators. They use measures of building height to street width ratios, and the presence/absence of sidewalks, parking areas, and vacancies, in order to give an indication of streetscape quality. These quality measurements are related to aspects of an idealised “active” street.

In Queensland, “active street frontages” to development are commonly related to conceptions of streetscape quality in both metropolitan policy as well local government land use regulations. There are few studies examining active frontages. Gehl et al. (2006) examine relationships between behaviour and active frontages, and Heffernan et al. (2013) study public perceptions of these spaces. Studies that attempt to quantify active frontages are even fewer. Kickert (2015) undertakes an impressive mapping exercise to quantify longitudinal changes to active frontages in two cities. However, this study assumes the presence or absence of active frontages based on land use. No existing studies could be found that quantify active frontages across urban areas based on direct observation of their presence. As observations of Google Street View were used to create a land use database, it was possible to simultaneously observe building frontage types when creating the database. Each lot was assigned one of six frontage types (Table 8). Where an active frontage was identified, the length was traced from aerial imagery, and historic GSV and aerial images were used to identify lengths of past active frontages. The lengths of active frontages can then be combined for each centre to compare changes over time. As different centres have different amounts of uses that could be expected to have active frontages, the active frontages are divided by an estimate of the length all possible frontages (the shape length of selected building footprints divided by four) in each centre.

Table 8 - Building frontage classifications

Frontage Type Description None No structure on the site Mall Commercial uses set far back form the street, or disengaged from the street through obstructions or off-street car parking areas Urban Active Small or no setback to a public street, with an active frontage featuring windows and openings along the frontage, and opportunities for interaction Urban Non-Active Small or no setback to a public street but with an access only and no active features Urban Residential Residential use with a small or no setback to a public street

Suburban Typical suburban streetscape characterised by large setbacks from a public street consisting of driveway/s and landscaping

Based on the above review of mixed-use measures, the following indicators have been selected to measure the mixed use aspect of compact activity centre policy:

• Land use variation (the Entropy Index of desired land use types)

• Average Euclidean distance (the average direct distance between different land uses)

• Median residential proximity (the average of the closest median distances between residential uses and other land use types)

• Proportional proximity of residential uses (the proportion of dwellings within 400m of commercial uses to all dwellings)

• Active frontages (the proportion of active frontage lengths to all commercial frontages)

These measures can all be considered at different points in time in both absolute and relative terms. However, like the issues associated with measuring changes to the IQV of housing types, the entropy index must also be considered with caution for this purpose. If increases to the index are considered as being an improvement, centres that were already well mixed at the earlier point in time and which maintained their mix, would not show significant improvement in the ENT score as it would have already been close to the maximum of 1.

Baseline measures The measures previously described allow for a highly detailed examination of how the identified centres have changed over time in-line with metropolitan policy. Also of interest is how these changes compare to changes in non-centre locations. If metropolitan planning policy intends to focus residential and employment growth within centres, then after twenty years there would be an expectation of higher rates of growth in the centres compared to locations outside the centres. Baseline measures (i.e. measures of developed areas outside of the centres) can offer an indication of these differences. The previously described indicators are designed specifically to measure centre policy, and most are therefore not suitable for measurements of the broader conurbation. Data availability also limits suitable baseline measures. As previously discussed, issues around the poor availability of detailed land use data for centres were resolved through manual Google Street View observations. The use of GSV was feasible for the centres due to the relatively small areas involved. Undertaking the same manual observations across the broader conurbation however is not practical and purchasing commercial land use data exceeds the research project’s available resources.

100

It is also important that the baseline measures allow for comparisons of relative change. Apart from the “new” centres of North Lakes and Springfield, in 1996 all the centres consisted primarily of existing urban uses. Baseline comparisons therefore need to minimise comparisons to changes occurring on previously undeveloped areas. A map of 1991 existing urban areas was obtained from the regional planning policy documents to identify existing urban areas (The State of Queensland, 1993c, p. 13). This map was georeferenced in ArcMap, and the outline of existing urban areas within the 35km study area radius, was traced to identify already developed areas.

Based on available data the following baseline measures were developed. Each of these figures can be considered in terms of the entire conurbation, or inner, middle and outer areas as previously described. The calculations use the ring based approach used by Coffee, et al. (2016) and as described as follows:

Density • Relative population and dwelling change. This measure is based on the relative change between summations of the non-centre areas of 1996 collection districts and 2016 SA1 areas that fall within the existing urban area. As previously discussed, the geographies used in the census have changed over time. The overall areas of comparison therefore differ between years. To account for these changes, the 2016 figures were therefore divided by the 2016 land area, and then multiplied by the 1996 land area. This is a form of area weighted interpolation, which is considered appropriate in this instance due to the large geographies and absolute figures involved.

• Population and dwelling density by built-up area. This measure divides the population and dwelling numbers calculated previously by the area of the cadastral boundaries contained by the census geographic units, and also contained within the “urban residential” and “services” (excluding defence and recreation) use categories defined in Queensland land use classifications9. The services category is required to capture land associated with some mixed used developments which are recorded under this category in the land use dataset. To ensure centre areas are measured against the same criteria, new density figures are also calculated for each centre using this same lot selection method. As the Queensland land use data lacks any further precision in residential classifications, it is not possible to accurately interpolate residential uses further. The census data outside of the centres may therefore include population and

9 https://data.qld.gov.au/dataset/land-use-mapping-series/resource/7c91ddea-84d4-4935-8134-c31228b6184c

101

dwelling numbers from any rural properties that happen to be within a selected census geographic area. The baseline density figures may therefore be slightly higher than the actual density for the selected lots. The same process is used for the 1996 densities, however the land areas are selected using the 1999 Queensland land use classification mapping. The mapping is the closest available to the census date of 1996.

• Proportion of population living at low densities. This figure has been calculated using the method of baseline population calculations described previously, where the population living in low density dwelling types are divided by the total population.

Diversity of dwelling types • All the measures of dwelling diversity can be determined at the baseline level and compared to the centres. Dwelling numbers for each category of dwelling are calculated using the same method for overall dwelling numbers described previously.

Employment • A lack of appropriate land use data prevents the calculation of employment density. The other indicators used to measure centre employment change are also not possible to calculate as baseline figures as these are based on building footprint areas which have not been created outside the centre areas. Employment can therefore only be compared in terms of relative change of the baseline employment numbers. The baseline employment figures were calculated in the same way as the baseline population and dwelling numbers except using Journey to Work data with 1996 SLAs and 2016 DZNs. Terrill et al. (2018) note that a change of methodology in 2016 census JTW job allocations resulted in data at smaller geographies being underestimated compared to 2016 census data, and propose a method to overcome this issue. Their method (Terrill, et al., 2018, p. 60) was used to determine an inflation factor of 1.086 for the greater Brisbane area, and this was multiplied by the 1996 employment totals to make them comparable to the 2016 figures.

Mixed use • The lack of detailed land use data means no meaningful comparison can be made between the developed measures of centre mixed-use and the broader urban area.

Analysis The research question for Objective 1 seeks to understand how activity centres have changed in-line with compact city based metropolitan policy. To address this question, a data set of individual properties (n=44,063) and unique development sites (n=26,003) has been created, recording land use and built form changes over a 20 year period. Through aggregations of the

102 property level data for each centre, and the incorporation of secondary data, it has been possible to address a series of key indicators of compact activity centre conformance. These indicators can be compiled for each centre at the beginning and end of this 20 year period, and changes between centres directly compared. Changes can also be compared to baseline measures for urban areas outside of the centres to give an indication as to whether the centres are changing in a manner different to the broader conurbation. It is also possible to combine the indicators to create an overall indicator of centre conformance, relative to the other centres.

Although the source data set is extensive, for this particular objective the unit of analysis is the defined centres identified in metropolitan policy. There are a total of 21 centres, which include two new centres (Springfield and North Lakes) that cannot usefully be compared to the other centres in terms of intensification. Due to the relatively small number of centres, combining indicators into an overall index using techniques such as factor and principal component analysis (as undertaken by Ewing, et al., 2014) is not possible. The small number of centres however does permit the direct comparison between centres, and their change over time. Along with direct comparisons of the absolute values of each indicator, z-scores can be calculated to measure a combination of indicators (Burton, 2002; Galster et al., 2001; Stathakis & Tsilimigkas, 2014). This method is simple to calculate and understand. Z-scores are calculated by subtracting the mean of the score from the score for each centre and dividing by the standard deviation of the scores. The result is an expression of the value of the indicator in terms of number of standard deviations it is away from the mean. The score approaches zero the closer the value is to the mean. Where required, the z-scores can be inverted by multiplying by -1 to ensure that positive z-scores are indicative of centres that score “better” in terms of achieving metropolitan activity centre goals, than the average. The z-scores can be added or averaged together to combine the results of different indicators. To reduce the possibilities of any one indicator having an overly dominant effect, this study follows the approach used by Burton (2002) and averages the z-scores to develop rankings of each indicator category, as well as a total compactness ranking. A similar analysis is also undertaken on changes to each indicator over time to develop rankings of centre intensification. Although it is possible to assign weightings to the scores, in this instance there is a lack of supporting information on what such weightings ought to be. The scores have therefore been given equal weighting.

Conclusion Policies to create a network of compact activity centres in greater Brisbane have been in place now for more than two decades. This section has provided a methodology to evaluate

103 conformance with this policy by examining how activity centres have changed in-line with key indicators of centre compactness. To be in keeping with metropolitan policy intent, centres should display changes over time in terms of:

• Increases in residential densities;

• The development of a greater diversity of dwelling types;

• Increases in employment; and

• A greater mix of land uses.

However, measuring these aspects in discrete urban areas is complicated by poor data availability and precision. These issues were overcome through the use of direct observations of the built environment using Google Street View and aerial imagery to create a detailed database of built form and land use change, and combining this database with secondary sources such as census data, address counts, and government land use coverage data. A range of commonly used indicators of urban compactness were reviewed, and a selection of measures were identified or developed on the basis of data availability, and their suitability to measure compact activity centre policy. A summary of these measures, the data used, and the data sources, are shown in Table 9.

Table 9 - Summary of compact activity centre indicators

Indicator Description Data Source Density

Net population density Persons per net residential hectare Land use Google Street View (2016) Net dwelling density Dwellings per net residential hectare database Nearmap aerial imagery Average land area of Total net land area of low density (2016) low density dwellings residential uses divided by the number Google Earth historic aerial of low density dwellings imagery (2001-2016) Proportion of Population of low density uses divided QLD government historic population living at low by total centre population aerial imagery (1995-1997) densities ABS Census (2016) Census data ABS Census (1996)

Dwelling Mix Proportion of low The lower the proportion, the more Land use As above density dwellings compact the overall dwelling types database Census data As above

Proportion of low The higher the proportion, the more medium density compact the dwelling types dwellings, medium

104

Indicator Description Data Source density dwellings, and high density dwellings Index of Qualitative The closer to 1, the more mixed the Variation of dwelling dwelling types types Employment Net job density The number of jobs divided by the area Land use As above of employment land in hectares database

Average employment The floor area of employment buildings Building Google Street View intensity divided by the number of jobs footprints (2016) Nearmap aerial imagery Employment plot ratio The floor area of employment buildings (2016) divided by the area of employment land QLD government historic aerial imagery (1995-1997)

Workspace ratios City of Sydney floor area survey’s (2012)

Shopping centre Property Council of floor area data Australia (2016) Building Owners and Managers Association (1993)

Census data ABS Census – Place of work (2016)

Mixture of uses Land use variation The Entropy Index of desired land use Land use As above types database Average Euclidean The average direct distance between distance different land uses Median residential The average of the closest median Land use As above proximity distances between residential uses and database other land use types Building As above footprints Proportional proximity The proportion of dwellings within of residential uses 400m of commercial uses to all Address point Queensland Department of dwellings data Natural Resources and Mines (2016)

Active frontage The proportion of active frontage lengths to all commercial frontages

105

These indicators are then compared with each other, as well as with baseline indicators of changes in the broader Brisbane conurbation in order to determine the degree of conformance of actual centre change to metropolitan planning objectives. The results of this analysis are described in detail in Chapter 5.

4.2.3. Objective 2: Evaluate the performance of Brisbane’s regional scale policies for compact activity centres in informing land use regulations, and the conformance of actual land use changes to these regulations. Metropolitan planning for activity centres in greater Brisbane anticipates implementation to occur primarily through the regulatory land use planning system (see section 2.2). This system is predominantly administered by local governments who create localised strategic and regulatory plans, assess development proposals against these plans, and enforce compliance. Applying the framework by Faludi and Altes (1994, see Chapter 3 for further discussion), for the plan to perform local governments should be using the metropolitan plan when making decisions about zoning in centres to enable the types of uses proposed by the activity centre policy. Such decisions require significant commitments to a future land use pattern. Zoning decisions give rights in land, are legally binding, and reversing direction through the subsequent removals of rights can be politically and economically costly. Queensland planning legislation also requires public consultation to be included in the process of rezoning land, giving voice to a range of potential critics or supporters of proposed changes. Regulatory land use plans also formalise the framework for subsequent decision making on individual development applications.

On this basis, examining local government land use planning decisions is a useful approach to evaluate metropolitan policy for activity centres. As previously discussed, the approaches typically used in performance evaluations involve either interviewing/surveying key actors or undertaking some form of document analysis. In this instance, the study is investigating change over a 20 year timeframe, over an area that covers nine separate local governments (including defunct local governments) with sometimes changing jurisdictional boundaries. Identifying suitable actors from each of these local governments, who can specifically recall land use decisions made on discreet local areas over this time frame, is considered to be impractical for the scope of the research. Examining local government land use regulations and other relevant policies for evidence of performance offers a more practical approach. The limitation of this approach is that where the regulations do not reference the regional policy, it is not possible to determine if the regional policy was used and purposefully ignored (a performance success) or if it was not used in decision making at all (a performance failure)

106

(Faludi, 2006). However, in this instance this limitation is considered to be acceptable as it is hypothesised that such eventualities are expected to be rare. The review of regional policy in section 2.2 indicated that Queensland local governments have been supporters of the activity centre policy. The voluntary nature of the initial Regional Framework for Growth Management policies implies a high degree of local government “buy-in” to the policy. The subsequent South East Queensland Regional Plans have all been developed with local government input, even though the statutory nature of the regional plans now compel local governments to use the regional policy when making their own planning instruments and deciding development applications. It is therefore hypothesised that whether voluntarily, or forced by legislation, that local governments will have made use of the compact activity centre policy when formulating land use regulations.

This research firstly examines local government land use regulations for the presence of terminology and references to metropolitan planning policy. By examining the regulations in force in approximately 1996, 2006, and 2016, it is possible to identify how regional planning policy for activity centres has been incorporated over time. Changing land use regulations can be a complicated process that takes many years to complete. Ten year periods have therefore been selected to allow sufficient time for the regulations to catch-up to regional policy, and also to align with the key data sources used to determine land use change (see section 4.2.2). The text of the regulations is reviewed to determine if they specifically reference the relevant regional plan for the time period or if they incorporate regional planning terminology. Where regional planning terminology for activity centres is specifically referenced in the regulations, the regional plan is considered to have been used in decision making for the development and adoption of the regulations. A total of 26 regulatory documents were examined for evidence of regional policy for activity centres. The planning regulations often have complex histories of amendments. It is difficult to obtain a complete history of past regulations and all their various amendments, as records of these are often incomplete. Of the documents available, versions of the regulations dated as close as possible to the key dates of 1996, 2006 and 2016 were therefore used. A summary table of the documents used for each centre can be viewed in section 10.4. The documents were reviewed in terms of their overall strategies, as well as in terms of the provisions that are applicable to each of the activity centres being studied.

As local governments have been required to reflect the regional plan in land use regulations from 2005, there may be instances where the regional plan is referenced in regulations primarily as a form of lip-service to the regional policy, and the actual land use regulations display little evidence of change. Ultimately, it is the detailed land use regulations and their

107 application to individual properties through zones and local plans that determines their material effects.

As discussed previously (section 2.2), metropolitan scale polices for compact activity centres intend to create a network of centres characterised by higher density residential uses, a greater diversity of housing types, and mixed clusters of uses that generate employment and provide localised services. As the principal implementation mechanism for this policy, land use regulations should reflect these objectives. Where local governments have used compact activity centre policy as a part of their land use regulations, the regulations would be expected to change over time to be more permissive of higher density residential uses and more intensive retail and office uses. Land use regulations should also show changes that permit a greater variety of uses on individual properties. Metropolitan policy also seeks to locate significant industrial uses outside of centres. Performing land use regulations would therefore be expected to show changes which reduce the amount of land available for these uses. Regulatory plans that exhibit these forms of change over time, combined with the incorporation of activity centre terminology, are considered to be positively performing.

The intensity of development permitted by the land use regulations affecting the centre extents can be quantified to enable an assessment of the nature of these material changes to the regulations. Most studies that attempt to take measure of land use regulations do so using aggregated measures applied at the jurisdictional level (Gyourko & Molloy, 2015; Talen et al., 2016). The vast bulk of such studies occur in the field of property economics and are associated with correlating regulations with housing supply and prices, rather than the evaluation of plans (Chakraborty et al., 2009; Quigley & Rosenthal, 2005). Due to the complexities in directly measuring land use regulations across multiple jurisdictions, studies of this nature typically rely on surveys of planning authorities to create indices of regulatory stringency that are applied generally to the entire authority’s jurisdiction (Gyourko et al., 2008; Pendall et al., 2018). For this study, the relatively small centre extents require a fine-grained examination of land use regulations. These regulations can be considerably different across centres and within a single regulatory jurisdiction. Due to the 20 year timeframe involved, surveying the local governments to determine the permitted land uses is not possible. As it is possible to obtain past land use regulations, directly assessing these documents offers a practical way to measure the intensity of development permitted by the regulations.

Studies that do so are relatively rare and usually select key regulatory aspects for comparison such as minimum lot sizes or permitted numbers of dwellings/density (Evenson & Wheaton, 2003; Gabbe, 2018; Glaeser et al., 2006). Studies that measure commercial and other non- residential based regulations are even more uncommon (Evenson & Wheaton, 2003). The

108 lack of research in this area is based on the difficulty of obtaining time series data, and challenges in making comparisons between regulatory aspects that often differ in terms of terminologies and the aspects of development that are regulated (Dain, 2005; Gyourko & Molloy, 2015). The regulations themselves are also often structured differently by different jurisdictions and frequently require the interpretation of multiple, overlapping sets of development requirements (Dain, 2005). These aspects all apply in the case of land use regulations for greater Brisbane’s activity centres. Different local governments apply different combinations of zoning, strategic level plans, development control plans, and other regulations. The relative importance of these instruments also varies between local governments, requiring each jurisdiction to be considered individually. The use of “typical” development controls such as minimum lot or density requirements is also not possible for comparative purposes, as different local governments use different standards at different times. For example, the current Brisbane City Plan 2014 uses “form based” codes that generally do not stipulate maximum residential densities while for other jurisdictions, density is an important consideration. Others include density requirements in some areas but not in others, and these requirements come and go even within jurisdictions as planning schemes are amended or replaced. What is common however, are regulatory intents that specify desired land uses and their approximate scale. For residential uses, building heights are typically specified thereby creating a physical limitation on the scale of their development, while non- residential uses are typically limited in terms of gross floor areas, heights, and/or intended service area catchments. The interpretation of these aspects enabled the development of a scale of permitted development intensity for different use types.

To create this scale, planning scheme maps were obtained from libraries, local governments, city archives, and internet archives. Where necessary, they were georeferenced in ArcMap, and the relevant areas were traced to create electronic versions of the mapping. Where GIS based data could be sourced, it was directly imported into ArcMap. The planning scheme provisions for each relevant layer were then examined and assigned Development Intensity Scores (DIS) for each part of the planning scheme. A development intensity score was assigned for residential, commercial, industrial, and bulky good retail uses (Table 10). The various planning scheme layers were then combined in accordance with their relative importance as described in each planning scheme, to create an overall layer that captured the relevant regulations applicable to each centre in approximately 1996, 2006, and 2016. This layer enables a DIS to be applied to any property in the centre extent and therefore gives an indication of the permitted types of residential, commercial, industrial, and bulky goods retail uses on any site, at the selected points in time. It should be noted that these scores capture only the intended uses and the desired scale of development. They do not include site based

109 limitations such as building setbacks, open space provisions, car parking standards etc. These aspects are not considered to be important in terms of the research objective to test whether local governments are adjusting their planning regulations to accommodate the spectrum of uses proposed by regional policy for compact activity centres. Constraints, such as environmental areas and flooding, are also not included in the DIS score as the historical mapping for these aspects was not available for the majority of centres. This limitation is not considered to be significant as in most instances, local governments typically do not zone highly constrained sites for intensive development. The centres are also located in existing urban areas that have minimal constraints, and which have usually been subject to decades of past urban development.

The criteria in Table 10 have been derived from the terminologies and types of use classifications observed in planning schemes themselves. As described above, there are significant differences in how local governments establish thresholds for different types of development. The purpose here is to capture an overall planning intent for a site that is comparable across the various local government jurisdictions. It was therefore necessary to have a degree of flexibility in the DIS criteria to account for these variations. The scoring bands across use categories are independent of one another and are not used to make direct comparisons with different use categories. For example, a score of 2 residential use does not have any relationship to a score of 2 for a commercial use; the research considers these as entirely different forms of development. The number of scoring bands for each use category were derived from the differentiation in scales observed in the planning regulations themselves, for those use types. For example, commercial uses are typically restricted by their intended catchment, and physical scale, and the six bands of score accounted for the gamut of scales observed in the various planning scheme. In comparison, planning schemes had few gradations for bulky goods retail uses, and these use types could therefore adequately be captured in only three score bands.

Table 10 - Development Intensity Score (DIS) criteria

Score Residential (RES) Commercial (COM) Industry (IND) Bulky Goods Retail (BGR)

1 None None None None

2 Very low density use Individual, small Individual, small Bulky goods retailing – single dwellings on scale, convenience scale service that has a limited large lots uses catering to industries catering scale – local residents (e.g. local residents (e.g. general stores, cafe) repair shops etc.)

3 Low density Commercial uses, Industries with low Unlimited GFA or residential – including small impacts that are large scale dwelling houses and shopping centres limited in scale. Can that primarily cater include warehousing to the needs of a

110

duplexes (typical local catchment, or suburban style lots) offices – low rise or lower floors only if in a mixed use development.

4 Low-medium density Commercial uses, Industries with low N/A residential – including shopping to medium impacts primarily houses, centres, catering to and warehousing. and town houses local level Little or no limit on catchments, or scale offices – no limit on ground floor/podium

5 Medium density Commercial uses, Most industrial uses N/A residential – including shopping permitted with little primarily apartments centres catering to or no limit on scale up to three storeys sub-regional catchments. Permitted to occupy large proportion of the site (not just ground/lower floors), up to five storeys

6 Medium-High Commercial uses, N/A N/A density residential – including shopping apartments four to centres catering to six storeys sub-regional catchments. Permitted to occupy large proportion of the site (not just ground/lower floors), over five storeys

7 High density N/A N/A N/A residential – apartments seven or more storeys

To gauge the reliability of the DIS coding, a test-retest comparison was undertaken using Cohen’s κ, similar to the reliability testing undertaken for the GSV land use database described on page 84. A total of 1,322 planning scheme parts were initially coded with DIS scores. A random sample of 70 of these parts was determined to be sufficient to provide the desired statistical power and alpha (0.9 and 0.05 respectively) for testing κ (Cantor, 1996; Gamer, et al., 2019). The sampled sections were recoded more than a year after the initial coding. The sample coding was compared with the initial coding for agreement with the four DIS types: residential score, commercial score, industrial score, and bulky goods retail score. The results of the agreement tests are shown in Table 11 below which include κ and its confidence interval, the coefficient of raw agreement, and κn (Brennan & Prediger, 2016; von Eye & Mun, 2005). The results of Kendall’s W are also included to demonstrate the degree of concordance when the DIS are considered as ordinal data (Kendall, 1962; von Eye & Mun, 2005).

111

Table 11 - Results of reliability testing of DIS coding

Observation Cohen’s κ Raw agreement ra Brennan & Kendall’s W

type Prediger’s κn Residential 0.944 (95% CI, 0.883 to 1) 0.957 0.950 0.997 score Commercial 0.927 (95% CI, 0.858 to 0.992) 0.942 0.932 0.989 score Industrial 0.910 (95% CI, 0.825 to 0.994) 0.942 0.930 0.991 score Bulky goods 0.946 (95% CI, 0.873 to 1) 0.971 0.961 0.985 retail score

Using the categories for κ identified by Fleiss (1981), the coding demonstrated excellent agreement between tests and was statistically significant beyond chance agreement (p < 0.05).

The use of the DIS enables changes to the regulated development intensity to be tracked over time, in each centre. By applying the DIS to every property, this analysis can record the amount of land designated for different forms of development, including sites where a mix of uses are possible. Changes to the amount of land assigned to different development types are used to give an indication as to whether local governments have decided to incorporate policies for compact activity centres in their regulatory documents. As informed by the review of regional planning policy in section 2.2, planning schemes that reflect the regional policy for compact activity centres should show evidence of:

• Greater proportions of land allocated for higher density residential uses;

• Greater proportions of land allocated for more intensive commercial uses, including a provision for bulky goods retailing;

• Reduction in the proportion of land for larger scale industrial uses; and

• Greater proportions of land that permit both residential and commercial development (i.e. mixed use development)

The land areas for each DIS were summed by centre to quantify these changes and the centres were classified according to their degree of performance. Centres in which regulations that changed in-line with activity centre policy intent were considered to be positively performing. Centres were considered to have marginal positive performance when only a minor change to regulations was observed (change of <5% of centre area10), or where the regulatory changes

10 The average centre land area is 224 Ha which includes diverse mixes of land uses including open space areas. The 5% of total centre area on average represents approximately 10 Ha, which is considered to be an area sufficiently large for development of

112 did not show a clear picture of overall intensification/reduction. Centres that saw regulatory changes contrary to activity centre policy were recorded as policy exceptions.

This approach seeks to determine plan performance on the basis of the outputs of planning decisions when making or amending land use regulations. Where land use regulations specifically mention the regional policy, and the regulatory documents themselves can be shown to be more permissive of the uses expected, the policy is assumed to be performing well by deduction. This analysis addresses local government decisions in relation to using regional policy when making plans. However, local governments also make regular decisions on the application of land use regulations to individual development proposals to intensify or change land use. To address this aspect, the conformance of land use and development to planning regulations need to be considered. A deductive approach is also used here to link conformance to performance. This assumes that if development decisions are occurring in accordance with the plan, then the plan is being used in decision making and is therefore also performing.

It is common for studies of land use conformance to consider conformance to regulations from data sources that offer a binary test: e.g. conversion of agricultural land was planned/not planned; or planned use x, equals use change y (Alfasi, et al., 2012; Chapin, et al., 2008; Padeiro, 2016). This is typically due to the nature of the regulations being examined, which are often based on single class zonings, or due to plans and/or land use data sources being described by coarse use groupings that lack detail of the relative intensity of the desired/actual use. This type of approach is less suitable for evaluating activity centres which, in most instances, already contain a diverse mixture of uses at the commencement of the study period, and where land use regulations commonly permit a broad range of possible uses on a single site. The previously described land use database categorises all land in the study area by a large number of uses, and also includes physical details such as building heights. This dataset provides a detailed source of land use change over time. By combining the land use dataset with the DIS representing the regulations in place in 1996, 2006, and 2016, it is possible to test conformance of not only use types, but also relative use intensity, while allowing for regulations that are permissive of a range of different uses on a single site. For example, mixed- business/industry zones are a common feature, particularly in outer centres. The DIS for a site in such zone may allow for a small shopping centre, a small warehouse, light industry, or limited scale bulky goods retailing. A land use can therefore be compared against any of these scores to determine if it conforms to the regulatory intent.

significance. For example, the average increase in land area used for commercial uses between 1996 and 2016 was 2 Ha. The 5% figure is used for an initial classification purpose which is subsequently qualified by a detailed examination of the circumstances of each centres (see Chapter 6).

113

Loh (2011) argues that binary classifications of conformance/non-conformance are too simplistic and fail to account for non-conformances that are in fact a natural part of the property development cycle. For example, an existing house in an area zoned for commercial uses most likely represents a situation where the zoning intent has yet to be realised. At some point in the future, the house owner sells the property to a developer who constructs an office building, and the use would then be conformant. Loh contrasts this to situations where a new use directly conflicts with the planned intent – a true case of non-conformance. This approach is an important consideration for an evaluation of activity centres in Queensland, where the statutory planning system is aspirational; that is, regulations are best considered as the intended future use of the land rather than a reflection of existing use. Where land use conformance is considered at a single point of time, it is therefore difficult to assert that a non- conformance is problematic as it can simply be explained as an existing use that has yet to realise its development potential. All land in the study area was therefore first categorised as being conforming, under-developed, or exceeding.

Conforming areas are parcels where the existing land use aligns with a DIS that reflects the use. For example, a commercial use such as strip mall or main street would be conforming if located on a parcel that had DIS that is permissive of commercial uses of this nature (i.e. equal to 3 or greater), while a more substantial big box shopping centre would require a DIS of 5 or greater to reflect the larger catchment areas expected of these types of uses. Attached residential uses were given a latitude of one storey (either under or over the recorded regulations) to account for possible variations in how land use regulations were classified11. Under-developed areas are parcels where the scale of the existing land use is less than the scale of use intended by the land use regulations. Examples include detached dwellings in high density residential areas, or vacant land that is permissive of other uses. Exceeding areas are parcels where the existing land use is of a type or scale that conflicts or exceeds the land use regulations. For example, an industrial use located on a site that has a DIS that does not permit industry (i.e. equal to 1), or apartments in a location intended for detached dwellings.

It should be noted the classification only applies to the observed use and does not consider the potential for mixed use development on a single site. For example, most regulations for big box shopping centre sites (Commercial DIS 5 and above), will also permit high density residential uses on the same site. Where a big box centre exists without residences, the lack of

11 Planning schemes vary in how they regulate height. Some use meters, while others use storeys or a combination of the two. A height in storeys is preferable as precise heights could not be determined from Google Street View observations. Counting the number of storeys in GSV images however is possible, and DIS based on storeys is therefore necessary to align with the GSV generated land use database. Where metres were used in a planning scheme, a height in storeys was determined using a standard storey height of 3.75m.

114 conformance with the residential component is not considered as the observed use meets the conformance criteria for the observed use (i.e. the criteria for big box centres). The full list of the conditions of classification for each use are shown in section 10.6.

The proportion of the land area for each type of use classification was then calculated for each centre in 1996 and 2016. These figures give an indication as to the amount of land in each centre where the use is intended to change, and can be considered as a possible factor influencing centre conformance as described in Chapter 7. This is similar to the approach used by Alfasi, et al. (2012) to measure changes in conformance over time. For this research however, the land use database records and locates every instance where a land use changed and can therefore isolate the changed uses for analysis. This enables analysis in terms of the number of planning decisions rather than in terms of land area. This is an important distinction as land area calculations can give distorted results in situations where a use change occurs over very large or small areas. Considering changes on the basis of individual development sites better reflects the actual processes of land use planning decisions which are made on discrete proposals for development. Under the Queensland planning system, all use changes must “comply” with planning regulations. In most instances, compliance is determined by local governments either through granting of a development permit for a specific proposed use, or through the creation of regulations that automatically permit certain forms of development provided they meet specific criteria. As this chapter primarily seeks to determine the degree to which local governments are making decisions in-line with compact activity centre policy, analysis of land use change using development sites is considered to best align with this objective.

Development sites were identified using the previously generated building footprints. Where applicable, these footprints were merged so that a single polygon represented a single site. The footprints were joined spatially to the land use database, and sites where the use changed between 1996 and 2006 were identified. In these locations, the previously created building footprints for 2016 substituted the 1996 footprints to create an approximation of the 2006 building footprints. Next, where a use changed between 1996 and 2006, the new use was classified for conformance using the 1996 DIS. Where a use changed between 2006 and 2016, it was classified using the 2006 DIS. This tests the conformance of a use change against the plan regulations that were most likely to have been in force at the time of the change. An exception was made for sites that changed between 2001 and 2006 in the Brisbane city area in order to account for the release of the Brisbane City Plan in the year 2000. These sites were classified against the 2006 DIS as this better reflects the City Plan 2000 provisions that were in place at the time of the change. The other local governments did not change their planning

115

schemes until closer to 2006. The number of conforming, under-developed, or exceeding sites for each centre were then counted for comparison, and to calculate the proportions of conforming and non-conforming uses for each centre. This approach also enables the calculation of the proportion of sites that were intended to change, but which did/did not change, as analysis relating to the types of uses that changed/conformed/did not conform.

Each changed use is reflective of a local government planning decision to enable the change. If a high degree of conformance is found, it therefore suggests that the local governments are using the regulations as are they are intended, and the regulations are performing well. Further analysis of non-conforming cases can consider the nature of these changed uses and the implications of these issues of non-conformance to the broader objectives to create more compact activity centres.

4.2.4. Objective 3: Identify and assess commonly cited factors that influence compact city progress. Section 4.2.2 describes a method to quantify the degree of compactness for each activity centre, as well as the extent to which the centres became more compact (or didn’t) over a 20 year period. The results of this analysis can be compared to a number of commonly cited explanatory factors for the poor implementation of compact activity centre policy to determine if there are relationships between these factors and policy implementation. In section 3.2.2, key themes were identified to explain the implementation of the compact city including demographic patterns and employment distribution (Birrell, et al., 2005; Troy, 1996), consumer preferences, infrastructure, and/or property economics (Bryant, 2013; O'Connor & Healy, 2004; Rowley & Phibbs, 2012; Searle, 2004, 2010), transport accessibility (Dodson, 2010), and political and institutional commitment (Brewer & Grant, 2015). In this section, a number of measures are proposed to quantify these factors and compare them against actual centre change. The measures for each factor are described and grouped into categories of existing compactness, property, transport, planning policy, and socio-economics status. Each measure was then tested for its correlation with compact centre intensification using Pearson or Spearman rank order correlation coefficients, controlling for the influence of other variables.

Intensification of centre compactness This measure ranks each centre on the basis of how much it has intensified in-line with regional planning objectives for compact activity centres. It was derived from a range of measures of changes to residential density, dwelling mix, and employment, in all activity centres between 1996 and 2016 (excluding the greenfield centres of Springfield and North Lakes). After testing the mixed use indicators, it was found that these indicators only recorded

116 very small changes in most centres. These aspects were therefore excluded from the overall intensification measure. The density component measures changes to population and dwelling numbers, net residential population and dwelling density, and areas of low density housing. The dwelling mix component considers the relative change of numbers of different types of dwelling, and the employment component measures changes to the amount of jobs, employment density, employment intensity, and the plot ratio of employment based buildings. The final intensification score is the average standard score (z-score) of these measures and is shown in Table 38 (p164). The method used to calculate the various components of intensification score are described in detail in section 4.2.2.

This overall intensification score is used as a quasi-dependent12 variable to consider which of the explanatory variables are related to its change, and which may therefore best explain the key influences relating to compact centre policy implementation.

Pre-Existing compactness In their critique of Melbourne’s activity centre policies, Birrell, et al. (2005) predict that the existing nature of the nominated centres (among other factors discussed later) is a key influence of whether the intended higher order employment and residential density will be realised. Under this view, only centres with sufficient services to “draw demand from a surrounding regional economy” are likely to develop as intended by urban consolidation policies.

The previously developed centre compactness scores rank each centre by how well they align with compact activity centre principles detailed in metropolitan policy. Comparing the compactness scores from 1996, to the degree to which each centre became more compact, can show if the existing nature of the centre is related to the achievement of activity centre policies. Compactness scores were determined for each centre on factors relating to density, dwelling mix, employment, and mixed land use. An overall compactness score was determined by averaging the standard scores (z-scores) of each of these factors. Full details of the various components of these scores and how they are calculated are described in section 4.2.2, and the scores for each centre are shown in Table 70 (p295). A short description of these variables is provided in Table 12.

12 The subsequent correlations are two way tests that do not differentiate between dependent and independent variables. The term of “quasi-dependent variable” refers to its role in being the target of the investigation and against which the other variables will be compared.

117

Table 12 - Existing compactness variables

Variable name Description Source 1996 Compact Score Average score of density, dwelling mix, employment and See below mixed use scores below 1996 Density Score Average of standard scores for the 1996 net residential See Table 9, (p104) population and dwelling density, average land area of low- density dwellings, and proportion of population living at low densities. 1996 Dwelling Mix Average of standard scores for 1996 proportions of low See Table 9, (p104) density, low-medium density, medium density, and high- density dwellings. 1996 Employment Score Average of standard scores for 1996 net employment See Table 9, (p104) density, employment intensity, the plot ratio of employment-based buildings. 1996 Mixed use score Average of standard scores for the 1996 land use See Table 9, (p104) variation, average Euclidean distance between use types, median residential distance to other use types, proportion of residential uses with 400m of commercial uses, and proportion of active frontage lengths to all commercial frontages.

Property factors Property economics and the role of consumer preferences for housing are commonly cited factors to explain the implementation (or not) of the compact city (Birrell, et al., 2005; Brewer & Grant, 2015; Bryant, 2013; O'Connor & Healy, 2004; Rowley & Phibbs, 2012; Searle, 2004, 2010; Troy, 1996). As described in section 2.2, the key implementation mechanism for compact activity centre policy is changing land use regulations to enable private development to instigate change. Change is therefore dependent on the potential for profit in undertaking the development. For higher density residential uses, construction costs represent the most significant expense in delivering this type of use (Birrell, et al., 2005). Such residential uses would therefore be more viable in locations where the sale price of the end product (units) are higher. Correlating measures derived from property values, with centre intensification, is undertaken to determine if there is a relationship between property price and centre development.

The most reliable sources of primary property price data in Australia is usually via government land title office, or state real estate institutes and their records of sale (Abelson & Chung, 2005). Unfortunately, title offices restrict and sell the release of point based property sales data making it costly to obtain records across large areas. This cost exceeds the resources available to the research. Suburb based medians and averages however are freely released and published on websites such as realestate.com.au. Although the suburb boundaries do not perfectly align

118 with activity centre extents, the suburb boundaries are considered to be an appropriate approximation for the centre’s overall residential property values. Median house and unit values from 2016 were obtained from the realestate.com.au website. Suburb based 1996 median house values were obtained from the Real Estate Institute of Queensland’s annual property review (REIQ, 1996). Unfortunately, median unit values were only published at the local government level. The 1996 unit price for each suburb was therefore estimated based on the proportional difference between the house and unit price listed for the relevant local government area, multiplied by the suburb based house price. These unit prices, and their absolute change between 1996 and 2016, make up the property variables as shown in Table 13. Unit prices were selected over house prices as they better represent the form of dwelling proposed by activity centre policy and the development processes to implement it. Historical commercial property prices could not be obtained.

Table 13 - Property variables

Variable name Description Source Unit Price 1996 Estimated unit price in 1996 REIQ 1996 Perct Unit Price Change Percentage change in median unit price between 1996 REIQ, 1996 and and 2016 realestate.com.au, 2017 Absolute Unit Price Difference in median unit price between 1996 and 2016 REIQ, 1996 and Change realestate.com.au, 2017

Transport factors Contrary to some expectations (Birrell, et al., 2005; Dodson, 2010), existing research suggests that levels of public transport accessibility bare little relationship to the actual occurrence of urban consolidation (Newton & Glackin, 2014; Phan, et al., 2009). By correlating measures of public transport accessibility with intensification of centre compactness, it is possible to test if this remains the case for greater Brisbane’s activity centres. The Spatial Network Analysis for Multimodal Urban Transport Systems tool (SNAMUTS) provides indicators of public transport accessibility across a range of Australian and international cities, including Brisbane (Curtis & Scheurer, 2015). SNAMUTS includes indicators for connectivity, population and employment catchments, network resilience, and transfers, and includes an overall composite index which combines its range of indicators (Curtis & Scheurer, 2015; SNAMUTS, 2016). The overall composite index for Brisbane is published in map format with accessibility measured on an eight interval scale ranging from “very good” to “no score”. The map was georeferenced in ArcMap and each centre was assigned the accessibility score located at its central transport node. Brisbane’s SNAMUTS scores are derived from 2011 data. As there

119 have been few material changes to the public transport network affecting the centres13 during this period, this is not considered problematic, and similarly rigorous measures of transit accessibility from the start of the study period could not be sourced.

Road distance was included to give some indication of motor vehicle use, and as a general variable of proximity. The transport variables are shown in Table 14.

Table 14 - Transport variables

Variable name Description Source SNAMUTS Rank from SNAMUTS composite index of public transport SNAMUTS, 2016 Composite accessibility Road Distance to Shortest distance by road from centre’s central node to CBD Google Maps, 2017 CBD

Planning policy factors This research is evaluating the implementation of compact activity policies on the basis of both plan conformance and performance. In order to undertake this analysis, aspects related to conformance and performance were quantified. The conformance components make up the quasi-dependent variable, against which other factors were correlated. The research has used observed changes in local government land use regulations to determine if these changes aligned with regional policy intent and therefore determine if local governments are implementing the activity centre policy as intended (Section 4.2.3 and Chapter 6). The intensity and type of development permitted by land use regulations were recorded across all sites in the study area at the beginning and end of the study period. The degree of conformance of actual land use change to land use regulations, and of existing land uses to land use regulations, were also determined for each centre. Brewer and Grant (2015) note that “institutional commitment” is one of the key impediments to increased residential densities in suburban areas. Local governments have adopted regional policy differently in different centres. “Committed” local governments would be expected to seek to increase or decrease the amount of land available to different land use types and intensities to achieve more compact activity centres. Such changes are the key implementation mechanism of regional policy. The changes to these regulations over time and the degree of land use conformance to regulations, can be correlated against centre intensification.

Variables measuring changes to land use regulations have been devised based on changes to regulations for residential, commercial, industrial, and bulky goods retail, calculated as the

13 The only significant public transport upgrade affecting the centres was the South East Busway which linked Upper Mount Gravatt in 2001. The rail line to Redcliffe was not completed until late 2016 which is outside the study’s timeframe. Springfield received a new rail line however as it is a greenfield centre it is not considered in this part of the study.

120 area adjusted difference in DIS between 1996 and 2016. The difference between each property’s grouped 2016 and 1996 DIS was multiplied by its shape area. These results were then summed by centre and normalised by the overall centre’s area. The degree of conformance is recorded as conforming, under conforming, or exceeding. As discussed in the previous chapter, exceeding sites are relatively rare. Conforming and under conforming locations will therefore be a ratio of mostly each other. The under conforming measure was therefore selected as this is theoretically of more interest; i.e. it represents the sites with some degree of regulatory development potential. The list of planning policy variables is shown in Table 15.

Table 15 - Planning policy variables

Variable name Description

Changes to land use regulations between 1996 and 2016

Change in residential zoning Summed difference in grouped residential DIS multiplied by property area, and intensity - 1996 to 2016 normalised by centre area. Change in commercial zoning Summed difference in grouped commercial DIS multiplied by property area, intensity - 1996 to 2016 and normalised by centre area. Change in industrial zoning Summed difference in grouped industrial DIS multiplied by property area, and intensity - 1996 to 2016 normalised by centre area. Change in bulky goods retail Summed difference in grouped bulky goods retail DIS multiplied by property zoning intensity - 1996 to 2016 area, and normalised by centre area.

Conformance of 1996 land use to 1996 land use regulations and development activity

Prop 1996 Under developed Sites Proportion of sites in each centre where the land use in 1996 was under developed in relation to the type/intensity intended by 1996 land use regulations Adoption of regional policy in 1996 1996 Strategic Reference The type of reference to regional policy found in planning regulations in 1996 (see section 6.1).

Socio-economic factors There is some evidence that the probability of land use conversion is partly based on socio- economic status (Kline & Alig, 1999; Padeiro, 2014, 2016), and that there is a relationship between high income households and the age of housing stock (Brueckner & Rosenthal, 2009). Socio-economic factors have therefore been included to determine if a relationship exists between the initial socio-economic level of the centres, and their degree of intensification. The Australia Bureau of Statistics (ABS) publishes a series of Socio-Economic Indexes for Areas (SEIFA). The SEIFA indices for 1996 were obtained from the ABS website. The indices are best considered in quantiles, as ordinal data (ABS, 1998). An index of each centre was determined by selecting the 1996 Collection Districts that intersect the centre extents and joining the SEIFA data to these boundaries. The SEIFA scores from this selection were ordered into deciles and the SEIFA scores for the centres were determined using the “population weighting” method as recommended by the ABS (ABS, 2008, p. 19). The

121 population weighted SEIFA score was then matched to the corresponding decile to create a rank ordered variable. The various SEIFA indices are highly correlated with one another which complicates their subsequent interpretation with the other variables. Of the four indices, the Index of Education and Occupation was selected as it was the only measure that was not directly composed of aspects related to the other variables (ABS, 1998) (for example, the inclusion of dwelling size measures in the index would create a bias for detached housing) and is described in Table 16.

Table 16 - Socio-economic variables

Variable Description Source 1996 IEO The decile rank of the 1996 Index of Education and Occupation – measures ABS 1996 the educational and occupational structure of communities on issues such as occupational status and educational attainment.

Analysis The land use database consists of more than 24,000 records of unique development sites and represents all land parcels within all of the nominated activity centre extents in the greater Brisbane area. However, as the overall unit of analysis is the activity centres themselves, these records are compiled to report on a total of 21 centres. Two of these centres (North Lakes and Springfield) are greenfield centres and therefore cannot be reliably compared to the other centres in terms of intensification. This leaves a total of 19 centres for analysis. Ideally, some form of multiple or logistic regression would be performed in order to predict how the dependent variable (centre intensification) changed based on the various independent variables. However, with only 19 cases and a large number of independent variables, this form of statistical modelling is unlikely to produce reliable results if there is an objective to apply its results to a wider population (Maxwell, 2000; Peduzzi et al., 1996). To undertake such research, a larger number of centres would need to be included in the analysis which would require extending the study to other Australian capital cities. This is not possible in this study as it is beyond its scope and available resources. The small number of centres therefore limits the type of statistical analysis that can be undertaken. Two common statistical approaches used to examine the relationship between variables are the Pearson Product-Moment Correlation and the Spearman Rank-Order Correlation (Corder & Foreman, 2014).

The Pearson correlation tests whether a linear relation is present between two continuous variables. It assumes there is a linear relationship between the variables, that they are normally

122 distributed, and not-skewed (Corder & Foreman, 2014; Myers et al., 2010)14. These aspects were investigated in SPPS using a visual interpretation of scatter plots, calculations for kurtosis and skewness, and running Kolmogorov– Smirnov and Shapiro-Wilk tests of normalcy. Of the groups of variables presented above, only the existing compactness group met all the conditions of the Pearson correlation for all variables. The other groups included variables that either did not meet the relevant tests or included ordinal variables. In these situations, non-parametric statistics, such as Spearman correlations, can be used instead (Corder & Foreman, 2014; Siegel & Castellan, 1988). Rather than testing a linear association, Spearman correlations covert the variable scores to ranks (essentially making them ordinal) and testing whether there is a monotonic relationship between the variables.

Both Pearson and Spearman correlations result in an r value on a scale of -1 to +1, where -1 represents a “perfect” negative relationship between variables, +1 is a perfect positive relationship between the variables, and 0 represents no relationship. Using SPSS, Pearson correlations were undertaken between existing compactness variables, while Spearman correlations were performed on all other variables. Statistical significance (p) for each correlation was also calculated in SPSS, which compares the r value to the critical r level for α = 0.05. Where r exceeds the critical value, there is a 95% chance that the observed correlation would not equal zero in a wider population and can therefore be considered as a “real” relationship that is not due to chance (Corder & Foreman, 2014). Reporting p is a standard procedure for sample based data. However, in this instance, all principal and major activity centres in greater Brisbane are included. The observed correlations therefore represent observations of the entire population and reporting of statistical significance is technically not required. However, some consideration must also be made as to whether there will be some inference to other areas associated with the study. For example, if the results are used to suggest that policy makers should consider alternative locations for future activity centres based on its conclusions, or use the results as an explanation of activity centre change in other cities or locations, then standards for inferential statistics should be considered. It can also be argued there will always be a degree of error associated with any study of a population (Abadie et al., 2014), or that a given finite population is actually part of infinite “super population” (Lavrakas, 2008), and that using statistical process to consider these aspects is therefore always of relevance. The study therefore reports significance levels for the various correlations, even

14 …although a normal distribution is not relevant if the centres are considered as a population rather than as a sample of a population – see discussion below)

123 though the correlations are used in a descriptive sense and are the real correlations observed for all activity centres in greater Brisbane.

The potential variation (PV (r2)) can be calculated from the correlation scores to determine what percentage of the relationship can potentially be attributed to the interaction of the two variables. Determining the exact strength of a correlation on the basis of its r value is subjective and dependent on the nature of the study, with a number of different classification scales or “rules of thumb” being proposed (Cohen, 1988; Corder & Foreman, 2014; Evans, 1996; Hinklle et al., 2003). In this instance, five key groups of factors have been drawn from the literature. If each of these factors contributed equally and without cross-correlations, a potential variation of approximately 20% (r = ~0.447) would account for the observed phenomenon. Observed correlations approaching and exceeding this figure are therefore considered and discussed, and correlated with each other to determine the strength of relationships that exist between the variables. Where cross-correlation between variables is strong, it introduces concerns about how much of the relationship with the dependent variable is the result of the cross-correlation. With only 19 cases in the analysis, multiple regression modelling with the required number of variables does not give stable results and is further complicated by the presence of ordinal variables (which also ought not be considered as continuous such as the SEIFA Index). Instead, the use non-parametric partial correlation is undertaken to indicate the strength of correlation while controlling for other variables (de Vries, 1993; Reynolds, 1974). These partial correlations were undertaken using SPSS software, where Spearman rank order correlations for the variables were outputted to a matrix and then partial correlations were calculated using the standard SPSS partial correlation procedure.

It should be noted that this approach of comparing correlations between variables is limited. Correlation does not equal causation, and the conclusions reached can only show that a relationship exists, not that the independent variable causes centre intensification. There may be the presence of confounding factors not included in the analysis which explain the phenomenon better than the selected variables. Additionally, the low statistical power associated with 19 observations limits the extent to which the relationships could be reliably inferred to a wider population (although in the case of statistically significant relationships, there is a greater than 95% chance that there is some relationship in a larger population). The primary purpose of the analysis is undertaken on a descriptive basis to evaluate the selected centres as a part of a case-study of the greater Brisbane area. It is therefore necessarily the role of future research, that expands the analysis to include additional centres, to test for causation and to permit inference to other places. The results of this analysis are detailed in Chapter 7.

124

5. Policy intent and reality: conformance of 20 years of metropolitan compact activity centre policy in greater Brisbane

This chapter details the results of research that seeks to evaluate how greater Brisbane’s activity centres have changed (or haven’t) in accordance with regional scale policies for compact activity centres over the past 20 years. The evaluation takes a conformance based approach that measures planned outcomes for development in each of these centres, against actual land use change. A review of two decades of metropolitan level policy for compact activity centres (section 2.2) revealed that a similar set of policies have been in place during this period. The policies seek to create more compact centres by:

• increasing residential densities;

• encouraging a greater diversity of dwellings types; and

• generating mixed-use clusters of business uses to provide employment and services to the centre catchments.

A series of indicators were developed to measure changes to these aspects by combining direct observations of the built environment using Google Street View and aerial photography, with a variety of secondary data sources such as census data (section 4.2.2). The indicators are considered in terms of their results from the beginning (1996) and end (2016) of the study period, and in terms of relative intensification between these years. Where possible, the indicators were also compared against baseline measures of change in the wider conurbation to compare whether centre change was the result of concentrated intensification of development in line with centre policy, or more a reflection of broader development trends. The results from each category of indicator are described in detail in the following sections of this chapter, before they are combined and assigned one of five categories based on a consideration of the centre’s characteristics of development, and their relative change in comparison the baseline changes.

The results demonstrate that conformance with regional planning policies has been poor, with few centres showing evidence of intensification that is consistent with all aspects of the activity centre policies or displaying evidence of growth that exceeds general development increases in the wider urban area. The results also highlight the considerable differences between centres, and a pattern of change that sees clusters of conforming centres in inner and middle locations, with non-conforming centres mostly located in outer areas. The lack of conformance was primarily the result of limited change, rather than of changes that were

125

directly contrary to compact activity centre policy. These results indicate that the activity centre policy is failing to be implemented as intended.

5.1. Conformance with objectives for higher residential densities

Proposals to increase residential densities in proximity to centres have been a consistent feature of regional policy, and key to the justifications of the purported sustainability benefits of more compact urban forms. The results show that although there is typically positive conformance in terms of reductions in the amount of land used for low density dwellings, the key indicators for population, dwelling and density change reveal that the majority of centres failed to intensify greater than baseline population and dwelling growth.

The following results are based on the selected density indicators (Table 17) which were assembled to evaluate centre conformance with policies for higher residential densities (see section 4.2.2 for details).

Table 17 - Indicators to measure residential density objectives

Indicator Description Data Source Density

Net population density Persons per net residential hectare Land use Google Street View (2016) Net dwelling density Dwellings per net residential hectare database Nearmap aerial imagery Average land area of Total net land area of low density residential (2016) low density dwellings uses divided by the number of low density Google Earth historic aerial dwellings imagery (2001-2016) Proportion of Population of low density uses divided by QLD government historic population living at low total centre population aerial imagery (1995-1997) densities Census data ABS Census (2016) ABS Census (1996)

5.1.1. Population and dwelling density The relative change to population and dwelling densities are shown in Figure 1115. There are clearly significant differences between how centres have densified. A pattern emerges of inner and middle centres showing greater intensification than outer centres, although there are some exceptions such as Carindale’s small increase in population density, and the larger changes shown in the outer centres of Capalaba and Cleveland. Different patterns of demographic

15 The 1996 and 2016 net population and dwelling densities, and absolute population and dwelling numbers are provided in section 10.2.1

126 change are also evident with some centres adding greater proportions of residents per dwelling (Logan Central and Beenleigh), while others have fewer residents per dwelling (such as Cleveland and Capalaba). Such differences most likely relate to the nature of new dwelling types, such as the retirement focussed apartment buildings of Cleveland and Carindale which are more likely to have fewer people per dwelling. The results of relative population, dwelling, and density changes are combined in the final set of indicators to determine the overall density score (see section 5.1.3 below), enabling the centres to be compared to each across a number of different measures.

Figure 11 - Relative change in net population and dwelling densities, 1996 to 2016

It is also useful to understand centre intensification in the context of wider urban development. If the policy is working as intended, the relative change in population, dwellings and densities should be greater within centre locations, compared to non-centre areas. Such results would show that centres are becoming a focus of more intensive development. Where possible, baseline indicators of inner, middle, and outer non-centre areas have been developed to enable a comparison to the level of intensification in centres with non-centre areas (section 4.2.2). The results of this comparison are shown in Table 18.

127

Table 18 - Difference in percentage points between relative baseline and centre changes to population, dwellings and densities, 1996 to 2016

Difference between centre and baseline relative… …population …population …dwelling …dwelling Centre Name Location density change change density change change16 Toowong Inner 1.6 -2.0 0.6 -2.9 Carindale -12.1 2.5 -11.1 2.9 Chermside 54.3 55.7 58.7 60.1 Indooroopilly 26.6 20.1 33.4 26.9 Mitchelton 7.0 12.7 12.1 17.8 Middle Toombul 30.1 25.8 31.5 27.4 Upper Mount -0.3 -1.4 2.2 1.2 Gravatt Wynnum 0.0 -4.2 5.8 1.7 Central Beenleigh -16.7 -30.4 -1.2 -14.2 Browns Plains -31.0 -28.3 -25.9 -23.5 Capalaba -14.1 -4.7 1.4 11.4 Cleveland 29.1 47.3 44.6 63.4 Goodna -25.7 -32.7 -13.9 -20.2

Ipswich Outer -35.1 -34.9 -19.6 -18.8 Logan Central -14.5 -31.7 0.9 -15.6 Logan -33.3 -32.2 -17.8 -16.0 Hyperdome Redcliffe -25.6 -22.6 -10.1 -6.5 Springwood -34.2 -38.0 -18.7 -21.9 Strathpine 2.6 3.4 3.1 3.9

Legend exceeds location baseline by > 5 percentage points less than baseline location by < 5 percentage points similar to location baseline by +- 5 percentage points

Note: Table 55 (section 10.2.1) provides the relative change in baseline population, dwelling numbers, and densities from 1996 to 2016 for inner (0 to 5km from CBD), middle (5km to 15km from CBD) and outer (15km to 35km from CBD) locations.

These results are contrary to the intent for metropolitan centre policies to concentrate higher density residential development in centres. In total, only five of the nineteen centres exceed baseline measures for changes in population, dwelling and residential density by more than five percentage points. Although most inner and middle centres are growing at a similar rate as surrounding areas, only half of the centres (Chermside, Indooroopilly, Toombul, and Mitchelton) have growth that exceeds baseline measures by more than 5 percentage points.

16 The density change figures are calculated in terms of built-up hectares in order to be directly comparable to baseline density figures. They differ to the net density figures shown in Figure 11 and Table 53. See discussion on baselines in section 4.2.2 for further explanation.

128

The outer centres conform poorly on this metric. Only Cleveland outpaces baseline growth in all categories. Capalaba and Strathpine grow at rates similar to non-centre outer areas, although Capalaba does display greater residential density change. All other outer centres have population and dwelling change that is less than the outer area baselines.

5.1.2. Low density dwellings Low density dwellings are defined as detached dwellings and duplexes (see Table 6 (p88) for further details). Table 19 shows the average land area of low density dwellings in each centre in 1996 and 2016, and the relative change. Similar to figures of density, the average low density area will show intensification (in this instance a reduction is representative of intensification) when low density residential uses are converted to higher density residential uses, or to non-residential uses. Subdivision into new low density uses would also reduce this figure. All centres show some reductions, with Capalaba demonstrating significant relative change. This effect is primarily related to the conversion of a number large low density lots to low-medium density townhouses. Wynnum also shows a large relative change however this result is based on a combination of low-medium density conversions and subdivision of existing low density dwellings to create additional low density dwellings (Table 23). The greenfield developments at North Lakes and Springfield are interesting in that they have arisen from a “blank canvas” not constrained by existing uses and fragmented ownerships. The majority of the in-centre Springfield low density lots were developed from 2006 onwards and these lots display a lower average size compared to lots in the other centres. North Lakes has low density dwellings constructed during two different periods, with the lots created after 2012 having average areas of 315m2. The lots developed between 1996 and 2001 however, are more in-line with typical historic lots sizes. The overall average lot area for North Lakes is therefore higher. As discussed in section 4.2.2, a baseline measure could not be determined for this indicator.

Table 19 - Change in average land area per low density dwelling, 1996-2016 (order by relative change)

Location Centre Name 1996 2016 1996 to Average Average 2016 land area land area relative per LD per LD change dwelling dwelling (m2) (m2) Outer Capalaba 1112 786 -29.3% Middle Wynnum Central 690 602 -12.7% Inner Toowong 714 643 -10.0% Outer Cleveland 881 796 -9.7% Middle Indooroopilly 874 790 -9.6% Middle Mitchelton 711 645 -9.4% Middle Toombul 658 598 -9.1% Outer Logan Hyperdome 1468 1362 -7.2%

129

Location Centre Name 1996 2016 1996 to Average Average 2016 land area land area relative per LD per LD change dwelling dwelling (m2) (m2) Outer Beenleigh 920 864 -6.1% Outer Ipswich 831 795 -4.2% Middle Chermside 646 621 -3.8% Middle Upper Mount Gravatt 706 680 -3.7% Outer Browns Plains 953 924 -3.1% Outer Goodna 1241 1203 -3.1% Outer Redcliffe 663 642 -3.1% Outer Logan Central 806 784 -2.8% Outer Strathpine 692 678 -2.0% Outer Springwood 748 741 -0.9% Middle Carindale 731 725 -0.8% Outer North Lakes17 N/A 580 N/A Outer Springfield18 N/A 360 N/A

Most centres have also seen reductions in the proportion of the total population living in low density dwellings (Table 20). As with changes to densities, these changes are greater in the middle ring centres. Comparing these figures to measures of baseline areas however shows that in 1996 most centres already exhibited a smaller percentage of population living in low density dwellings compared to non-centre areas (section 5.2 provides further detail about existing dwelling mixes). The majority of centres also saw higher rates of change on this metric than baseline areas.

17 North Lakes and Springfield are greenfield centres and were undeveloped in 1996. 18

130

Table 20 - Change in the percentage of population living in low density dwellings, 1996-2016, including baseline areas

Location Centre Name 1996 2016 1996 to Percentage of Percentage of 2016 population population change in living in low living in low percentage density density points dwellings dwellings Inner Toowong 34% 24% -10.2 Carindale 89% 75% -14.1 Chermside 68% 31% -36.7 Indooroopilly 57% 39% -17.8 Middle Mitchelton 93% 81% -12.1 Toombul 49% 29% -19.8 Upper Mount Gravatt 88% 70% -18.5 Wynnum Central 85% 79% -5.5 Beenleigh 65% 55% -10.8 Browns Plains 98% 94% -4.1 Capalaba 49% 36% -12.7 Cleveland 74% 48% -25.2 Goodna 85% 86% 0.9 Ipswich 84% 79% -4.9 Outer Logan Central 70% 62% -6.7 Logan Hyperdome 86% 84% -1.7 North Lakes N/A 49% N/A Redcliffe 90% 84% -5.7 Springfield N/A 90% N/A Springwood 76% 74% -2.3 Strathpine 54% 35% -18.4

Inner 66% 49% -17.2 Middle 91% 83% -7.4 Baseline Outer 93% 88% -4.6 All 88% 80% -8.1

5.1.3. Combining density indicators Examining changes to population, dwelling growth and densities in the centres reveals a mixed picture of change. Half of the inner and middle centres exceed baseline densification, and the others have similar rates of densification to the broader urban area. The outer centres however show poor conformance on these measures, with only Cleveland showing greater intensification to the baseline in terms of both population and dwelling changes. In terms of low density dwellings, most of the centres show reductions in both land area, as well as proportions of the population living at low density.

131

By combining z-scores of each indicator in 2016 and 1996, it is possible to compare how the centres perform in relation to each other at the beginning and end of the study period. Additionally, the differences between the indicators and the relative change in population and dwellings has also been calculated as z-scores to provide a measure of intensification during this period. These results are summarised in Table 2119.

Table 21 - Average density score in 1996, 2016, and average density intensification score 1996 to 2016

Average 1996 density Z Score – Average 2016 density Z Score – Average density intensification ordered by 1996 score ordered by 2016 score score – ordered by score Centre 1996 Score 2016 Score Centre Centre Score Toowong 1.91 1.62 Toowong Chermside 1.74 Toombul 1.41 1.49 Toombul Toombul 1.13 Strathpine 0.83 1.15 Chermside Cleveland 1.05 Chermside 0.60 0.58 Indooroopilly Indooroopilly 0.83 Indooroopilly 0.39 0.36 Strathpine Toowong 0.79 Springwood 0.22 0.29 Capalaba Capalaba 0.76 Logan Central 0.20 0.04 Cleveland Mitchelton 0.13 Wynnum Central 0.02 -0.09 Upper Mount Gravatt Upper Mount Gravatt -0.01 Redcliffe -0.02 -0.09 Wynnum Central Strathpine -0.08 Capalaba -0.03 -0.18 Logan Central Wynnum Central -0.14 Beenleigh -0.07 -0.24 Mitchelton Carindale -0.36 Upper Mount Gravatt -0.09 -0.25 Beenleigh Beenleigh -0.40 Carindale -0.19 -0.33 Redcliffe Logan Central -0.58 Cleveland -0.31 -0.37 Springwood Redcliffe -0.64 Mitchelton -0.34 -0.38 Carindale Logan Hyperdome -0.71 Ipswich -0.46 -0.63 Ipswich Ipswich -0.78 Browns Plains -0.93 -1.10 Browns Plains Browns Plains -0.82 Goodna -1.40 -1.50 Goodna Goodna -0.88 Logan Hyperdome -1.74 -1.70 Logan Hyperdome Springwood -1.04

The overall density intensification scores are plotted spatially in Figure 12. Outer centres, particularly those in the south and west of the city, show below average levels of intensification compared to the other centres. These density scores are used in combination with the other compactness indicators to develop an overall compactness intensification score in section 5.5, where the results are discussed in more detail and used to create a typology of centre intensification.

19 The full list of z-scores for each indicator and centre are shown in Table 57 and Table 58 in section 10.2.1.

132

Figure 12 - Centre locations with density intensification scores

133

5.2. Conformance with objectives for a diversity of dwelling types

Regional planning policies for centres seek not only to increase residential densities, but also to improve “housing choice by delivering a mix of dwelling types” (The State of Queensland, 2017c, p. 40). Aspects of residential “choice” are typically related to housing affordability, accessibility, and lifestyle goals, and align with the social sphere of the overall sustainable development framework. The indicators shown in Table 22 have been developed to evaluate centre conformance with policies for dwelling mix (see section 4.2.2 for further details). The indicators consider changes to the proportions of dwelling types, and also their degree of mix.

Table 22 - Indicators of dwelling mix

Indicator Description Data Source Dwelling Mix Proportion of low The lower the proportion, the more Land use database Google Street View density dwellings compact the overall dwelling types (2016) Nearmap aerial imagery (2016) Proportion of low The higher the proportion, the more Google Earth historic medium density compact the dwelling types aerial imagery (2001- dwellings, medium 2016) density dwellings, and QLD government historic high density dwellings aerial imagery (1995- Index of Qualitative The closer to 1, the more mixed the 1997) Variation (IQV) of dwelling types

dwelling types Census data ABS Census (2016) ABS Census (1996)

5.2.1. Proportions of dwelling types For low density dwellings (detached houses and duplexes), dwelling numbers were determined by counting the number of building footprints for these uses (duplex counts were doubled to reflect two dwellings occupying a single footprint). The dwelling numbers for other residential uses are estimates derived from dasymetric areal interpolation of census data.

Figure 13 shows the proportion of different dwelling types for each location in 2016 and 1996, as well as baseline proportions for non-centre areas. The dwelling categories are defined depending on their built form as described in Table 6 (p88). What is firstly apparent is how the centre locations already had a far greater diversity of dwellings in 1996 compared to non- centre baseline locations. This is particularly pronounced in the middle and outer centres where more than 30% of dwellings were low-medium or medium density types in 1996. In terms of change, all locations displayed reductions in the proportion of low density dwellings. The middle centres display strong shifts to more diverse dwelling types, with larger increases

134 in medium and high density dwellings compared to the non-centre areas. In the outer centres it is a different situation. Although low-medium density dwelling types represent a significant proportion of the housing stock, far exceeding baseline proportions, the change to different dwelling types has been more modest. The outer centres do show larger increases in the proportion of medium and high density dwelling types compared to baseline changes, but medium and high density uses still only make up 5% of dwellings in outer areas in 2016.

Figure 13 - Proportions of dwelling types in and outside centres, 1996 & 2016, by location

There are of course differences between the individual centres20. The relative change for each dwelling type by centre is shown in Table 23. These figures show that almost all centres a saw a reduction in proportions of low density dwellings, or if low density dwellings increased, they did so at a rate lower than the baseline rate. In outer baseline areas for example, low density residential uses represent the single largest increase in dwelling types. The outer activity

20 Figure 55 and Table 59 (section 10.2.2) show the differences between the proportions for each centre.

135 centres however are not following this trend, with most centres showing reductions in the proportion of low density dwellings. The middle centres are similar, however Wynnum and Mitchelton do show higher proportions of low density dwellings due to a pattern of small scale, one into two subdivisions.

The baseline measures for low-medium density dwellings show that these forms of housing are common in middle and outer non-centre areas. The middle centres typically match or exceed baseline changes. In outer areas however only Capalaba, Cleveland and Strathpine outpace their surrounding areas.

The most pronounced differences between baseline and centre changes relate to medium and high density dwellings types. This is primarily evident in middle ring centres, where the changes typically far exceed the baseline changes. In outer locations however, these forms of dwellings are not common either within, or outside the centres. Only Beenleigh, Cleveland, and Redcliffe deliver a significant quantum (more than 10021) of medium/high density dwellings.

21 Absolute dwelling change figures are provided in Table 60 in section 10.2.2

136

Table 23 - Relative change in dwelling types, 1996-2016

Relative LD Relative LMD Relative MD Relative HD dwelling change change Dwelling change dwelling change

Location Centre Inner Toowong -3.6% -0.5% 17.8% 25.1% Carindale 0.9% 18.3% 0.0% 6.6% Chermside -14.5% 39.6% 4.1% 49.8% Indooroopilly -2.3% 5.3% 14.5% 25.8% Middle Mitchelton 6.9% 12.8% 12.4% 3.9% Toombul -5.7% 10.2% 20.6% 24.0% Upper Mount Gravatt -4.1% 7.0% 3.7% 15.3% Wynnum Central 7.1% 8.3% 3.2% 0.5% Beenleigh -3.3% -1.1% 8.0% 0.0% Browns Plains 0.5% 5.2% 0.0% 0.0% Capalaba -3.9% 31.6% 1.6% 0.0% Cleveland 4.3% 53.3% 13.1% 10.5% Goodna 1.6% 0.2% -0.5% 0.0% Outer Ipswich -8.5% 1.9% 3.3% 2.3% Logan Central -0.4% 1.6% 0.0% 1.0% Logan Hyperdome 1.2% 0.6% 0.0% 0.0% Redcliffe -0.6% -0.4% 3.1% 9.2% Springwood -1.7% -2.4% 0.0% 0.0% Strathpine -0.6% 36.2% 1.8% 0.0%

Inner -3.8% 1.9% 9.4% 25.4% Middle 7.1% 9.2% 1.9% 1.6% Baseline Outer 20.9% 8.8% 0.4% 0.9% All 10.3% 7.6% 2.5% 5.4%

Legend

exceeds baseline by > 5 percentage points less than baseline by < 5 percentage points similar to baseline by +- 5 percentage points

Note: The low density dwelling column is coded inversely to reflect the normative position that more compact centres should see reductions of low density dwelling types.

Like the density measures, indicators for dwelling mix show uneven results, but higher degrees of conformance in middle locations. Where dwellings are added in any significant number, they typically are of higher density types. The issue for the outer centres however is the overall lack of development activity in general and several of the outer centres have seen little in the way of residential development. This is especially pronounced in Browns Plains, Goodna, Ipswich, Logan Hyperdome, and Springwood, where few new dwellings have been added.

137

5.2.2. Index of Qualitative Variation Applying a calculation for the Index of Qualitative Variation (IQV) to the proportions in Table 59, measures the degree of mixture between the dwelling categories, where 1 is “perfectly” mixed (i.e. the same proportions in each category) and 0 is not mixed (Table 24). The bulk of the centres were more mixed than the surrounding areas in both 1996 and in 2016. During this period, the middle centres have all either approached 1 or have seen changes similar or greater than the baseline IQV changes. The outer centres show an overall limited amount of change in terms of mix. Only Cleveland and Redcliffe had a dwelling mix that increased significantly more than baseline growth. Capalaba and Strathpine actually showed reduced scores on this indicator. This was due to the relatively large increases in low-medium density dwellings types in these centres, which were already a dominant form of dwelling in 1996.

Table 24 - Index of Qualitative Variation of dwelling types, 2016 and 1996

Location Centre 1996 IQV 2016 IQV Inner Toowong 0.99 0.97 Carindale 0.35 0.62 Chermside 0.76 0.96 Indooroopilly 0.88 0.99 Middle Mitchelton 0.26 0.62 Toombul 0.88 0.98 Upper Mount Gravatt 0.40 0.72 Wynnum Central 0.53 0.62 Beenleigh 0.67 0.76 Browns Plains 0.09 0.20 Capalaba 0.61 0.52 Cleveland 0.64 0.82 Goodna 0.44 0.42 Outer Ipswich 0.50 0.63 Logan Central 0.66 0.68 Logan Hyperdome 0.48 0.48 Redcliffe 0.35 0.54 Springwood 0.60 0.59 Strathpine 0.65 0.58

Inner 0.80 0.96 Middle 0.33 0.50 Baseline Outer 0.26 0.37 All 0.41 0.58

138

5.2.3. Combining dwelling mix indicators Table 25 displays the z-scores of each indicator in 1996 and 2016 to compare the centres relative to each other. Numerous outer centres have more mixed dwelling structures than several of the relatively low density dwelling dominated middle centres such as Carindale and Upper Mount Gravatt. However, this was also the case in 1996 where these centres already had significant stocks of low-medium density dwellings. These patterns therefore better reflect historic development rather than change to the centres during the study period.

The dwelling mix intensification scores are the z-scores of the relative change in the number of each dwelling type (i.e. the results shown in Table 23). These scores are shown alongside the centre rank scores in Table 2522, and plotted spatially in Figure 14. The pattern of intensification is similar to that of density intensification, with middle ring centres forming a cluster of above average conformance, and outer areas showing below average conformance. In centres where more than 100 new dwellings have been developed, the majority of the additional dwellings are of attached forms, adding to the diversity of dwelling types in each centre. Wynnum is an exception to this rule however and although more than 200 attached dwellings were developed, an additional 200 detached houses were also added. As this centre is characterised by a relatively large proportion of existing low density dwellings, the addition of large numbers of low density dwellings caused the centre to score poorly despite the development of low-medium density housing. The scores from Table 25 are combined to form overall compactness scores and discussed further in section 5.5, where they are used to contribute to a typology of centre intensification.

22 The full list of z-scores for each indicator and centre are shown in Table 61 and Table 62 in section 10.2.2.

139

Table 25 - Dwelling mix indicators for 2016 and 1996, and average dwelling mix intensification score

Average dwelling mix Average 1996 dwelling mix Z Score Average 2016 dwelling mix Z Score intensification Z Score – – ordered by 1996 score – ordered by 2016 score ordered by score

Centre 1996 Score 2016 Score Centre Centre Score

Toowong 1.90 1.42 Toowong Chermside 1.77

Toombul 0.98 1.20 Toombul Toombul 1.02

Indooroopilly 0.94 1.07 Chermside Toowong 0.67

Capalaba 0.59 1.01 Indooroopilly Cleveland 0.66

Strathpine 0.54 0.47 Cleveland Indooroopilly 0.58

Chermside 0.34 0.33 Capalaba Capalaba 0.11

Beenleigh 0.34 0.32 Strathpine Upper Mount Gravatt 0.10

Logan Central 0.21 0.22 Beenleigh Strathpine 0.02

Cleveland 0.03 -0.09 Logan Central Ipswich -0.02

Springwood -0.08 -0.11 Upper Mount Gravatt Beenleigh -0.19

Wynnum Central -0.31 -0.35 Ipswich Mitchelton -0.24

Ipswich -0.37 -0.38 Carindale Carindale -0.27

Logan Hyperdome -0.42 -0.38 Mitchelton Redcliffe -0.32

Goodna -0.53 -0.38 Wynnum Central Logan Central -0.57 Upper Mount Gravatt -0.61 -0.43 Springwood Browns Plains -0.58

Redcliffe -0.71 -0.53 Redcliffe Springwood -0.59

Carindale -0.72 -0.71 Logan Hyperdome Logan Hyperdome -0.68

Mitchelton -0.89 -0.83 Goodna Goodna -0.73

Browns Plains -1.22 -1.27 Browns Plains Wynnum Central -0.73

140

Figure 14 - Centre locations with dwelling mix intensification scores

141

5.3. Conformance with objectives for employment

Employment is considered to be a critical component of the activity centre policy. The intention of the policy is to increase employment, particularly in terms of retail, office, and entertainment uses. The policy reasons that doing so will provide more localised employment opportunities, and a more diverse array of services to the local community which will in turn reduce vehicle trips for both work and services and create employment efficiencies through clustering of business uses. The indicators shown in Table 26 have been developed to evaluate centre conformance with policies for employment (see section 4.2.2 for further details).

Table 26 - Indicators of employment

Indicator Description Data Source Employment Net job density The number of jobs divided by the Land use database Google Street View (2016) area of employment land in hectares Nearmap aerial imagery (2016) Google Earth historic aerial imagery (2001-2016) QLD government historic aerial imagery (1995-1997)

Average employment The floor area of employment Building footprints Google Street View intensity buildings divided by the number of (2016) jobs Nearmap aerial imagery (2016) QLD government historic aerial imagery (1995-1997)

Employment plot ratio The floor area of employment Workspace ratios City of Sydney floor area buildings divided by the area of survey’s (2012) employment land Shopping centre Property Council of floor area data Australia (2016) Building Owners and Managers Association (1993)

Census data ABS Census – Place of work (2016)

5.3.1. Job density The net job density indicator is based on employment estimates generated by applying known workspace ratios to estimations of floor area for various employment uses. As such, it is best

142 considered as a measure of a centre’s built form capacity for employment, rather than a measure of actual economic activity. The development of this method is described in detail in section 4.2.2. Employment land areas were determined through direct observations of the built form using Google Street View and aerial imagery.

Figure 15 shows the net employment density by centre for 2016 and 199623. This measure shows very large differences between certain centres. Carindale for example, is notable for its high density which is almost entirely the result of a large scale expansion to Westfield Carindale. Although Carindale has lower absolute employment than Westfield Chermside, the Carindale shopping centre is located on a smaller lot, with less external land area used for car parking, and it therefore achieves a higher overall density. Other big box dominated centres (Indooroopilly and Upper Mount Gravatt) also show high employment densities, and Chermside’s big box centre is bolstered by significant expansions to the nearby Prince Charles Hospital. Ipswich is noticeable among the outer centres both in terms of current density, as well as overall change from 1996. The city centre has seen the development of Riverlink Shopping Centre (a big box shopping mall), expansions to the Ipswich hospital, and a new multistorey office development led by a council owned development corporation (Ipswich City Properties – discussed further in Chapter 8).

All centres have seen some increase in employment density, except for Browns Plains. This centre has had a large number of new employment generating developments, including a new mixed use “main street”. However, it has also seen the development of land hungry uses within the centre such as large scale bulky goods retail, low intensity mixed industry development, and significant areas of warehousing. These uses are characterised by expanses of car parking combined with large floor areas and low job ratios, which have resulted in the reduction in overall employment density despite the addition of new employment. Toombul is another centre of interest and has seen the conversion of a number of employment uses into mixed-use residential towers. However, these towers often contain ground floor employment based uses that are small and which sometimes resulted in reduced employment capacity compared to the retail and light industrial uses they replaced.

23 Available in tabular format in Table 63, section 10.2.3

143

Figure 15 - Net employment density in 1996 and 2016

5.3.2. Employment in non-centre areas The measures developed for employment are reliant on detailed land use and building data. This type of data was not available for all non-centre locations in the greater Brisbane area. Comparative baseline employment measures for non-centre areas were only possible in terms of relative employment change (see section 4.2.2). Table 27 shows the absolute numbers of estimated jobs in 1996 and 2016, and the relative change between these figures. The results are colour coded to show how they compare to the baseline changes in employment for their respective locations. In the inner and middle centres, it is the big box shopping dominated centres that show increases greater than the baseline. In the outer centres, only Browns Plains and Cleveland show larger employment changes than baseline growth. With only six centres showing higher than baseline growth, the policy is conforming poorly in terms of objectives to focus employment based uses within centres.

144

Table 27 - Estimated employment and relative change, 1996 to 2016

2016 Est. 1996 Est. Relative Location Centre Employment Employment change Inner Toowong 8153 6355 28.3% Carindale 4941 1997 147.4% Chermside 13809 7958 73.5% Indooroopilly 6412 4032 59.0% Middle Mitchelton 3337 2919 14.3% Toombul 3879 3619 7.2% Upper Mount Gravatt 8258 4778 72.8% Wynnum Central 2464 2317 6.3% Beenleigh 3603 2559 40.8% Browns Plains 6964 2854 144.0% Capalaba 5856 4504 30.0% Cleveland 3839 2245 71.0% Goodna 1288 909 41.7% Ipswich 12719 8156 55.9% Outer Logan Central 3543 2836 24.9% Logan Hyperdome 4568 2913 56.8% North Lakes 5777 N/A N/A Redcliffe 4727 3289 43.7% Springfield 4031 N/A N/A Springwood 5619 4942 13.7% Strathpine 4994 3544 40.9%

Relative Baseline Legend baseline location change exceeds baseline employment rates Inner 35.5%

less than baseline employment rates Middle 44.9%

Outer 65.0%

All areas 42.2%

Some care must be taken when comparing the centre estimates to baseline figures. The baseline figures represent a response to the census questions of “for the main job held last week, what was the person's workplace address?” (ABS, 2016). The census data is therefore vulnerable to changes in levels of unemployment, as well as changes to the labour market such as changes in levels of temporary and casualised work. The estimates for the centres are based purely on changes to the built form and assumes occupancy and use at the level of the applied employment ratio. The two measures (although well related when tested against 2016 data (section 4.2.2)), are fundamentally different in these regards. Unfortunately, the census data for 1996 is only available at the SLA level, and it was therefore not possible to validate the 1996 centre estimates against the census figures (the SLAs are typically too large). None the

145

less, even if a large variation (say, +/- 20%) in the baseline data is assumed, the situation remains the same in that few centres would significantly exceed baseline growth.

5.3.3. Types of employment change The type of employment change that is occurring is typically in line with regional planning intentions. The employment change is occurring in desired use types such as shopping centres (i.e. combination of retail, office, and some services), institutions (hospitals, schools, etc.), and offices (Table 64, page 291). Some of the outer centres show growth in bulky goods retail. Bulky goods retailing can be problematic in that it makes use of large floor spaces, with relatively low employment rates, and often requires the movement of goods that are of size that require private transportation. These traits do not align well with overall compactness objectives but bulky goods retailing is none the less provided for in regional activity centre policy, and is therefore technically not contrary to planned intentions. Browns Plains was the only centre that showed large increases in uses not intended by centre policy (warehousing). Employment change shows differences between inner, middle and outer areas (Figure 16). Big box shopping centres offer the most significant type of employment growth across all areas, however in the inner and middle areas this use type dominates employment growth. Outer area employment growth is split between greater varieties of use types. Without shopping centre employment growth, the employment figures for the centres would be considerably more marginal compared to the baseline. Although shopping centres consist of uses typically desired for centres, the employment associated with their growth is likely to be retail focussed, and the big box nature of these developments tends to cater primarily to those arriving by car. This begs the question of how beneficial such dominant growth in this sector is to the broader sustainability objectives that justify compact activity centres, and whether planning policy needs to better consider the nature of employment that should be delivered in centres

146

Figure 16 - Percentage estimated employment change by use, top 5, 1996-2016

5.3.4. Employment intensity The average employment intensity indicator is based on the work place ratios as applied to the estimated floor areas and estimates the average floor area per employee in each centre. Using this indicator, centres that have adopted uses characterised by higher employee rates per m2, will show reductions in the amount of floor space per employee (Table 28). The pattern here is mostly of better (i.e. lower scores) in the middle and inner centres. There are some notable outer centres however. Redcliffe has a particularly good score, however this is explained by hospital uses dominating its overall employment (65%). Ipswich offers a diverse range of employment uses in a traditional town centre. These uses are typically space efficient, and its key employment groups of hospitals, shopping centres, and office boosts its scores in the measure. At the other end of the scale, the introduction of large areas of low employment generating uses such as bulky goods retail and warehousing has reduced the overall

147

employment intensity in Browns Plains; the only centre to see major falls. Otherwise, there is little differences between the centre’s 1996 and 2016 figures.

Table 28 - Average employment intensity by centre, 1996 and 2016

Employment Employment Location Centre intensity 1996 intensity 2016 Inner Toowong 32.96 31.48 Carindale 33.44 32.49 Chermside 24.65 25.05 Indooroopilly 40.26 40.37

Mitchelton Middle 42.44 41.38 Toombul 36.07 36.17 Upper Mount Gravatt 35.59 34.72 Wynnum Central 45.37 45.20 Beenleigh 39.03 38.68 Browns Plains 39.57 52.47 Capalaba 42.58 44.98 Cleveland 43.88 42.01 Goodna 41.61 41.17 Ipswich 35.57 34.75

Logan Central Outer 37.46 38.27 Logan Hyperdome 37.49 37.88 North Lakes N/A 34.23 Redcliffe 30.51 29.61 Springfield N/A 37.59 Springwood 47.95 47.80 Strathpine 42.41 42.07

5.3.5. Employment plot ratio Employment plot ratio gives an indication of the efficiency of the use of employment land. Higher plot ratios mean there is more useable floor area as a proportion of employment land. Higher plot ratios are typically generated through multi-story uses that maximise land use. Figure 17 shows the overall plot ratio for each centre in 1996 and 2016, and the difference between these figures24. Once again, the dominance of Westfield Carindale and its large floor area on a relatively small site (for a big box centre of its scale) shows the highest plot ratio. Other centres that had large scale big box shopping centre upgrades also scored highly. Toowong’s high rise built form yields higher plot ratios, as do the multi-storey offices of Ipswich. Toombul’s change to a more high rise built form did not translate to notably higher

24 Show in tabular format in Table 65, section 10.2.3

148 plot ratios as these new developments are primarily residential, and the associated employment uses are confined to portions of the lower floors.

Figure 17 - Employment plot ratio, 1996 and 2016

1.1.1. Combining employment indicators The combined z-scores for the employment indicators, and total intensification scores for each centre relative to the other centres, are shown in Table 2925. When comparing centre ranks, the middle centres are again mostly clustered at the top of the table. Wynnum lacks big box centres of the scale seen in other centres, as well as large scale institutions such as metropolitan hospitals, and therefore proved to be an exception to this rule. Mitchelton’s employment was almost entirely made up of a single shopping centre on a sprawling site, and its schools, also giving it a low score. Ipswich’s administrative focus, hospital, and new shopping centre make it exceptional in terms of employment for outer centres.

25 The full list of z-scores for each indicator and centre are shown in Table 66 and Table 67 in section 10.2.3

149

The intensification scores (plotted spatially in Figure 18) are based on the difference between the 1996 and 2016 net job density and employment plot ratio figures, as well as relative employment change. The average employment intensity was dropped from this measure as most centres showed very little change during the study period on this metric. The pattern of employment intensification is still characterised by the highest conforming centres in middle locations, however the average and below average centres are not as clustered by location. The scores from Table 29 are combined to form overall compactness scores in section 5.5, where they are discussed further and used to develop a centre intensification typology.

Table 29 - Overall employment scores for 1996 and 2016, and employment intensification score

Average employment Average 1996 employment Average 1996 employment Z Score intensification Z Score – ordered Z Score – ordered by 1996 score – ordered by 2016 score by score Centre 1996 Score 2016 Score Centre Centre Score Toowong 2.00 1.87 Carindale Carindale 2.81 Chermside 1.02 1.59 Toowong Chermside 1.07 Upper Mount Gravatt 0.54 1.43 Chermside Upper Mount Gravatt 0.87 Carindale 0.52 0.91 Ipswich Ipswich 0.60 Capalaba 0.07 0.90 Upper Mount Gravatt Indooroopilly 0.53 Redcliffe 0.03 0.26 Indooroopilly Browns Plains 0.35 Indooroopilly -0.10 0.15 Redcliffe Toowong 0.21 Logan Hyperdome -0.09 0.05 Toombul Cleveland 0.19 Browns Plains -0.08 -0.02 Logan Hyperdome Logan Hyperdome 0.06 Goodna -1.51 -0.41 Cleveland Redcliffe -0.16 Ipswich 0.72 -0.42 Capalaba Strathpine -0.36 Toombul 0.71 -0.47 Logan Central Beenleigh -0.49 Cleveland -0.79 -0.52 Wynnum Central Goodna -0.50 Logan Central -0.41 -0.61 Springwood Logan Central -0.56 Wynnum Central -0.24 -0.61 Strathpine Capalaba -0.57 Springwood -0.23 -0.69 Beenleigh Mitchelton -0.77 Strathpine -0.68 -0.73 Mitchelton Wynnum Central -0.79 Beenleigh -0.76 -1.00 Browns Plains Springwood -0.79 Mitchelton -0.73 -1.16 Goodna Toombul -0.94

150

Figure 18 - Centre locations with employment intensification scores

151

5.4. Conformance with objectives for a mixture of uses and accessibility

Creating more mixed-use centres is a key objective of regional planning policy as well as of conceptions of the compact city more generally. Similar to employment based objectives, it is reasoned that this type of built form will create better accessibility to services and reduce the need for vehicle trips. The activity centre policy also includes aspects related to streetscape improvements. The indicators shown in Table 30 have been developed to evaluate centre conformance with policies for mixed use (see section 4.2.2 for further details). Due to a lack of detailed land use data in non-centre areas, it has not been possible to make direct comparisons between these indicators and the broader urban area using baseline measures. This section therefore only makes comparisons between the centres.

Table 30 - Indicators of mixed use and accessibility

Indicator Description Data Source Mixture of uses Land use variation The Entropy Index of desired land use Land use Google Street View (2016) types database Nearmap aerial imagery Average Euclidean The average direct distance between (2016) distance different land uses Google Earth historic aerial imagery (2001-2016) QLD government historic aerial imagery (1995-1997)

Median residential The average of the closest median Land use As above proximity distances between residential uses and database other land use types Building Google Street View footprints (2016) Proportional proximity The proportion of dwellings within 400m Nearmap aerial imagery of residential uses of commercial uses to all dwellings (2016) QLD government historic aerial imagery (1995-1997)

Active frontage The proportion of active frontage lengths Address Queensland Department of to all commercial frontages point data Natural Resources and Mines (2016)

5.4.1. Land use variation The land use variation indicator applies a commonly used entropy index formula to calculate the degree of mix between different use types. Where the uses are perfectly mixed (i.e. the proportion of each category is equal) the score is 1, and it equals 0 when uses are not mixed

152 at all. Land parcels from the land use database were categorised as residential, retail (excluding bulky goods retail), office, or community uses, and the area of each of category was then summed by centre.

The results for 1996 and 2016 are shown in Table 31. The results show the outer centres averaging higher rates of variation than the inner/middle centres, and also exhibiting the largest difference between scores for 2016 and 1996. However, as shown by the previous indicators, actual intensification of land use in the form of residential development and employment is typically lower than the middle and inner areas. Most of the commercial intensification in middle areas, is occurring through redevelopment of existing land uses. Carindale for example, which scored highly on changes to employment density due to significant floor area increases on its existing shopping centre site), scores poorly on this measure as its commercial uses are highly concentrated and are surrounded by an expanse of low density dwellings with few other commercial uses, and little new commercial land was added. The outer centres however, often have larger areas of vacant or undeveloped land on which new uses can be established. In the inner/middle areas only 26ha of land changed from undeveloped (i.e. open space, vacant etc.) to other uses. The outer centres in contrast, saw 152ha of undeveloped land developed for other purposes, even when excluding the greenfield centres of Springfield and North Lakes. This therefore resulted in an increase in the amount of land that is factored into the land use variation calculation, and the score rises accordingly. Ipswich for example, although already displaying high land use variation, added approximately 14ha of retail use through the redevelopment of part of its railyards and gasworks to a new big box shopping centre.

153

Table 31 - Land use variation, 1996 and 2016

Location Centre Land use variation 1996 Land use variation 2016 Inner Toowong 0.55 0.57 Carindale 0.48 0.47 Chermside 0.75 0.73 Indooroopilly 0.59 0.60 Middle Mitchelton 0.58 0.56 Toombul 0.56 0.56 Upper Mount Gravatt 0.70 0.72 Wynnum Central 0.39 0.39 Beenleigh 0.70 0.77 Browns Plains 0.46 0.59 Capalaba 0.65 0.72 Cleveland 0.62 0.66 Goodna 0.44 0.47 Ipswich 0.81 0.90 Outer Logan Central 0.62 0.64 Logan Hyperdome 0.40 0.47 North Lakes N/A 0.85 Redcliffe 0.43 0.46 Springfield N/A 0.85 Springwood 0.72 0.76 Strathpine 0.76 0.76

Figure 19 shows differences in the proportions of land use mix between the centres in 1996 and 2016. In most instances, the proportions of the various types of land use have changed only subtly (Table 32). The notable exception of course can be seen in the increased proportions of higher density residential uses (townhouses and apartments), particularly in the middle centres. Of interest here is the trade-off between different land uses over time. For example, compare the higher density residential growth in Toombul with Cleveland. In Toombul, the change can mostly be accounted for through the conversion of low density residential land, while Cleveland’s change largely involved the development of previously undeveloped land. Areas with large movements between proportions also typically had correspondingly large falls in categories of undeveloped land.

154

Figure 19 - Proportions of estimated land use by centre, 2016 and 1996

155

Table 32 - Difference in proportions of area of estimated land use from 1996 to 2016

ail

etc.)

Retail

Centre

density)

Industrial

Shopping Shopping

Mixed Use Mixed

Residential Residential

Community

(open space, space, (open

Bulky Goods Goods Bulky

vacant, rural, rural, vacant,

Undeveloped Undeveloped

(higher density) (higher Residential (low (low Residential Location Centre Name Commercial/Ret

Inner Toowong -0.3% -0.6% 0.0% 0.3% -1.1% -0.4% 0.3% 9.2% -7.5%

Carindale 0.8% 0.0% 0.3% 0.0% 0.0% -3.7% 0.0% 2.5% 0.1%

Chermside -0.5% 0.6% 0.2% -3.4% -0.1% 0.2% 0.5% 9.6% -7.1%

Indooroopilly 0.4% -0.2% 0.1% 0.0% -0.2% 0.3% 0.5% 5.2% -6.1%

Middle Mitchelton 0.0% 0.1% 0.5% -2.1% -0.1% -0.8% 0.1% 3.2% -0.9%

Toombul 0.1% -0.7% 0.1% 0.2% -0.1% -0.1% 0.6% 7.8% -7.8%

Upper Mount Gravatt 0.9% 0.6% 0.1% 0.0% 0.0% -1.4% 0.2% 3.3% -3.6%

Wynnum Central 0.0% -1.1% 0.0% 0.0% 0.0% 0.4% 0.0% 2.7% -2.1%

Beenleigh 0.0% 0.7% 2.8% 0.4% 0.1% -2.3% 0.0% 2.2% -3.9%

Browns Plains 8.6% 2.6% 0.4% 0.2% 13.6% -25.3% 0.0% 0.5% -0.7%

Capalaba 1.7% 1.4% 0.1% 0.6% 2.0% -3.7% 0.0% 2.8% -5.0%

Cleveland 0.9% 0.3% 2.6% 0.7% -1.1% -10.7% 0.3% 8.0% -1.0%

Goodna 0.0% 1.0% 0.1% 0.3% -0.4% -1.0% 0.0% 0.6% -0.5%

Outer Ipswich 0.9% 1.5% 5.2% 0.0% -1.7% -3.0% 0.1% 0.6% -3.7%

Logan Central 0.0% 0.2% -0.5% 1.3% 0.0% -0.9% 0.1% 1.4% -1.6%

Logan Hyperdome 1.7% 1.4% 0.3% 0.4% 0.0% 0.2% 0.0% -0.3% -3.8%

Redcliffe -0.5% 0.3% 0.2% 0.1% 0.5% -0.2% 0.0% 1.5% -2.0%

Springwood 0.0% 1.0% 0.1% 0.3% -0.3% -0.6% 0.0% 0.4% -0.9%

Strathpine 0.0% 0.8% 1.4% 0.3% 1.5% -9.8% 0.0% 6.2% -0.4%

Cells highlighted in green indicate locations that have above average amounts of change for each use category

5.4.2. Proximity based indicators As discussed in section 4.2.2, measures of land use variance do not tell the full picture of land use mix as they lack a spatial component. Mixed use objectives require both a literal mix of uses and for the uses to be proximate to one another. The average Euclidean distance indicator measures the average distances between different use types, from a grid of points across the built-up area. This distance was calculated between residential, commercial (office and retail), and community uses for each centre (Table 33). The inner and middle centres show slightly more proximate uses on this measure, however the outer centres show greater changes between 1996 and 2016. This is again related to the development of new uses on undeveloped land. The difference here compared to the entropy measures is that there is now a spatial component. For example, Browns Plains saw the development of a new main street shopping

156 centre in an area close to its geographic centre, resulting in the third largest reduction in this distance measure. In contrast, the traditional town centre of Ipswich already had a wide variety of proximate uses in 1996. The addition of a new shopping centre in Ipswich however had less effect on this measure, as the new development was located on the centre’s fringe. Loganholme showed changes due to the nature of the uses themselves; a new library and police station, at opposite ends of the centre, reduced distances to community uses, thus resulting in the largest overall reduction in the average distance.

Table 33 - Average Euclidean Distance, 2016 and 1996

Euclidean Euclidean Location Centre distance (m) distance (m) 1996 2016 Inner Toowong 398 362 Carindale 854 762 Chermside 475 461 Indooroopilly 509 520 Middle Mitchelton 501 474 Toombul 426 409 Upper Mount Gravatt 564 546 Wynnum Central 504 502 Beenleigh 487 443 Browns Plains 1083 817 Capalaba 987 679 Cleveland 547 488 Goodna 606 559 Ipswich 355 330 Outer Logan Central 552 534 Logan Hyperdome 1114 769 North Lakes N/A 610 Redcliffe 543 525 Springfield26 N/A 1218 Springwood 562 502 Strathpine 617 551

The median residential proximity indicator shows the distance from residential uses to other uses. Unlike the Euclidean distance measurement of direct line distance, the median residential proximity indicator uses the actual road/path distance. This is calculated from each dwelling to the nearest park, commercial use (retail or office), or community use. Related to this measure is the indicator for proportional proximity of residential uses. Using the same

26 Springfield is still being developed. In 2016, no residential uses were visible in the centre core. This distance therefore reflects the distance to the low density residential uses on the fringe of the centre.

157

data, the measure records the proportion of dwellings that are within 400m walking distance of a commercial use. These two indicators are combined in Table 34. The key patterns observable in these results relate to both location and predominate use types. For example, centres characterised by big box shopping with large areas of predominantly low density residential uses such as Carindale, Upper Mount Gravatt, Browns Plains, and Logan Hyperdome, see correspondingly low scores on these measures. Ipswich however, as a traditional administrative and retail centre, shows the highest score. These indicators show little change between 2016 and 1996.

Table 34 - Indicators of residential proximity, 2016 and 1996

residential residential proximity 1996 Median residential proximity 2016 Proportional proximity res < (%dwelling 1996 400m) Proportional proximity res < (%dwelling 2016 400m) Location Centre Median Inner Toowong 315.30 309.61 83.6% 80.7% Carindale 541.51 523.86 37.2% 42.5% Chermside 363.58 342.06 71.4% 76.5% Indooroopilly 285.38 269.48 73.6% 76.4%

Middle Mitchelton 356.47 351.61 56.3% 57.7% Toombul 410.30 389.80 78.9% 83.6% Upper Mount Gravatt 545.95 553.55 31.1% 37.0% Wynnum Central 339.71 336.77 77.3% 76.5% Beenleigh 547.62 533.02 42.1% 44.5% Browns Plains 544.62 545.91 28.8% 29.1% Capalaba 591.49 519.21 10.4% 19.7% Cleveland 418.71 387.18 47.0% 61.1% Goodna 439.11 369.77 48.3% 53.8% Ipswich 255.27 254.35 89.9% 90.7%

Outer Logan Central 349.69 339.27 49.8% 50.9% Logan Hyperdome 623.11 603.29 32.4% 40.0% North Lakes N/A 482.91 N/A 37.9% Redcliffe 397.88 385.30 55.0% 58.2% Springfield N/A 879.17 N/A 0.0% Springwood 465.73 452.22 65.3% 67.8% Strathpine 533.72 519.68 43.2% 43.2%

5.4.3. Active street frontages The active frontage indicator is the final measure for mixed use. This is a design based measure created specifically to address regional planning objectives for more vital streetscapes. The planning objectives typically describe this in terms of a variety of uses that create an attractive pedestrian environment and therefore encourage more trips on foot. Measuring the extent of

158 street frontages that were “active”, was selected to quantify this concept. Through observation from Google Street View and aerial images, each property was assigned a frontage type. Properties with active frontages were then traced to collect data on the frontage length (Figure 20). Only frontages to public roads and spaces were included.

The centres show significant variation in the quantity of active frontages (Figure 21). Historic town centres such as Beenleigh, Wynnum, and Ipswich show extensive active frontages. Browns plains saw a considerable addition in active frontages through the development of its new “main street” shopping area. Chermside and Upper Mount Gravatt saw additions primarily in the form of ground floor mixed-use developments, while locations where big box shopping centres form the principal commercial use (such as Carindale and Logan Hyperdome) show almost no active frontages.

Figure 20 - Example of active frontage "tracing" in Logan Central (image source: NearMap 2016)

In order to make comparisons between the disparate centre types, the active frontage indicator is measured as the proportion of the active frontage length to an approximation of all commercial frontages within a centre. The results for 2016 and 1996 are shown in Table 35. On the whole, middle ring centres show greater proportional increases in active frontages compared to development in the outer rings. Some centres have seen falls from 1996, even though the total active frontage length increased. This is explained by new development not increasing active frontages at the same rate as non-active frontages.

159

Figure 21 - Active frontage lengths (metres), 2016 and 1996

0 500 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500

Toowong Inner Carindale

Chermside

Indooroopilly

Mitchelton Middle Toombul

Upper Mount Gravatt

Wynnum Central

Beenleigh

Browns Plains

Capalaba

Cleveland

Goodna

Ipswich

Logan Central Outer Logan Hyperdome

North Lakes

Redcliffe

Springfield

Springwood

Strathpine

2016 Active Frontages 1996 Active Frontages

160

Table 35 - Active frontages as proportion of all commercial frontages, 2016 and 1996

Active Active Difference Location Centre Frontage 1996 frontage 2016 (percentage points) Inner Toowong 35.4% 37.9% 2.5 Carindale 12.1% 9.5% -2.6 Chermside 20.4% 33.4% 13.0 Indooroopilly 24.8% 27.1% 2.3

Middle Mitchelton 23.7% 22.1% -1.6 Toombul 36.6% 40.8% 4.3 Upper Mount Gravatt 7.8% 10.6% 2.7 Wynnum Central 41.4% 48.3% 7.0 Beenleigh 44.6% 38.1% -6.4 Browns Plains 0.0% 10.2% 10.2 Capalaba 16.2% 13.5% -2.7 Cleveland 62.6% 45.6% -17.0 Goodna 15.7% 13.9% -1.8 Ipswich 49.7% 42.2% -7.5

Outer Logan Central 24.8% 21.9% -2.9 Logan Hyperdome 1.8% 1.2% -0.7 North Lakes N/A 19.5% N/A Redcliffe 14.3% 13.2% -1.1 Springfield N/A 0.9% N/A Springwood 4.1% 6.5% 2.4 Strathpine 16.4% 13.7% -2.7

5.4.4. Combining mixed use indicators As with the other categories of indicator, the z-scores for the mixed use measures have been combined to provide an overall mixed use centre ranking for 1996 and 2016 (Table 36). On these measures, Ipswich scores significantly higher than the other centres. As an historic town, it was already well mixed and accessible as is clear by its equally high scores from 1996. The combination of mixed-use measures work well in identifying centres that primarily consist of big box shopping centres, surrounded by suburbia. These uses are all clustered at the bottom of the rankings, while the other centres are closely grouped by score and show little change from their 1996 positions.

Measures of mixed use intensification (with the exception of the active frontage indicator), had very little variance between centres. High mixed-use intensification scores were often skewed by centres that contained areas of undeveloped land. As these types of changes were relatively large compared to the changes in more mature centres, mixed use intensification scores would show high relative change, even in centres that continued to have an overall low level of mixed use in the standard indicators. Conversely, centres that had higher levels of

161 mixed use in 1996, and that have shown large scale changes in terms of employment and housing, only registered marginal changes in terms of overall mixed use. The mixed use indicator is therefore considered to have limited value in measuring intensification changes, and is more suitable as a measure of mixed-use at a given time. For reference, the full list of mixed-use ranking and intensification z-scores can be seen in Table 68 and Table 69 in section 10.2.4.

Table 36 - Overall mixed use scores for 1996 and 2016

Average 1996 mixed use Z Score Average 1996 mixed use Z Score – – ordered by 1996 score ordered by 2016 score

Centre 1996 Score 2016 Score Centre Ipswich 1.53 1.45 Ipswich Toowong 0.78 0.75 Toowong Chermside 0.62 0.73 Chermside Indooroopilly 0.56 0.64 Toombul Toombul 0.55 0.57 Cleveland Cleveland 0.53 0.50 Indooroopilly Wynnum Central 0.38 0.42 Wynnum Central Mitchelton 0.25 0.31 Beenleigh Logan Central 0.23 0.15 Logan Central Beenleigh 0.21 0.15 Mitchelton Springwood 0.06 0.13 Springwood Strathpine 0.03 0.02 North Lakes Redcliffe -0.22 -0.12 Strathpine Goodna -0.38 -0.20 Redcliffe Upper Mount Gravatt -0.38 -0.22 Goodna Carindale -0.86 -0.32 Upper Mount Gravatt Capalaba -1.02 -0.52 Capalaba Browns Plains -1.33 -0.80 Carindale Logan Hyperdome -1.53 -0.84 Browns Plains North Lakes N/A -1.06 Logan Hyperdome Springfield N/A -1.74 Springfield

5.5. Discussion of results

The previous sections present the results from a range of measures designed to test the progress of activity centres in conformance to the overall objectives of Brisbane’s metropolitan planning policies. These policies seek to create a network of compact activity centres across greater Brisbane, charactered by a mix of higher density dwellings supporting a diverse mix of employment generating uses. The indicators used have been applied specifically to measures these factors and answer the research question of how Brisbane’s activity centres

162 have changed in accordance with metropolitan policy. The results show that few of the designated centres have conformed to the policy in a holistic fashion. This result is consistent with the results from studies of centre policies in other Australian cities that suggest such policies have had limited effects on urban development. However, the results also reveal that the degree of intensification of centres varies significantly across the different categories of indicator. Although few centres conform across all indicators, several centres show conformance in terms of individual categories of change, such as employment or dwelling mix.

Table 37 combines the 1996 and 2016 category scores into an overall compactness score league table by averaging the component z-scores27. Examining the differences between scores shows that in most instances, the centres are highly resilient to change, with centres that were the most compact in 1996 typically remaining so in 2016.

Table 37 - Overall compactness scores, 1996 and 2016, and overall intensification score

Overall 1996 compactness score Overall 2016 compactness score (ordered by score) (ordered by score)

1996 Centre 2016 Score Centre Score Toowong 1.65 1.34 Toowong Toombul 0.91 1.10 Chermside Chermside 0.65 0.84 Toombul Indooroopilly 0.45 0.59 Indooroopilly Ipswich 0.36 0.35 Ipswich Strathpine 0.18 0.17 Cleveland Logan Central 0.06 0.10 Upper Mount Gravatt Springwood -0.01 0.08 Carindale Wynnum Central -0.04 -0.01 Strathpine Beenleigh -0.07 -0.08 Capalaba Capalaba -0.10 -0.10 Beenleigh Cleveland -0.14 -0.14 Wynnum Central Upper Mount Gravatt -0.14 -0.15 Logan Central Redcliffe -0.23 -0.23 Redcliffe Carindale -0.31 -0.30 Mitchelton Mitchelton -0.43 -0.32 Springwood Browns Plains -0.89 -0.87 Logan Hyperdome Logan Hyperdome -0.94 -0.93 Goodna Goodna -0.96 -1.05 Browns Plains

27 The component scores have been described in the previous sections of this chapter and for reference, are listed together in Table 70, section 10.2.5.

163

These scores provide a snapshot of how the centres compare to one another at the two points in time. To investigate plan conformance it is also necessary to understand the degree of intensification. The combined intensification scores for each centre are shown in Table 38. The colour coding in the table is based the score’s distance from the average for each indicator group,28 and graded as either below average (score <= -0.5), average (score > -0.5 and < 0.5), or above average (score >= 0.5).

Table 38 - Overall centre intensification scores, 1996 to 2016

intensification intensification score Density intensification score mix Dwelling intensification score Employment intensification score Location Centre Overall Middle Chermside 1.53 1.74 1.77 1.07 Middle Carindale 0.73 -0.36 -0.27 2.81 Middle Indooroopilly 0.65 0.83 0.58 0.53 Outer Cleveland 0.64 1.05 0.66 0.19 Inner Toowong 0.56 0.79 0.67 0.21 Middle Toombul 0.40 1.13 1.02 -0.94 Middle Upper Mount Gravatt 0.32 -0.01 0.10 0.87 Outer Capalaba 0.10 0.76 0.11 -0.57 Outer Ipswich -0.07 -0.78 -0.02 0.60 Outer Strathpine -0.14 -0.08 0.02 -0.36 Middle Mitchelton -0.29 0.13 -0.24 -0.77 Outer Browns Plains -0.35 -0.82 -0.58 0.35 Outer Beenleigh -0.36 -0.40 -0.19 -0.49 Outer Redcliffe -0.38 -0.64 -0.32 -0.16 Outer Logan Hyperdome -0.44 -0.71 -0.68 0.06 Middle Wynnum Central -0.55 -0.14 -0.73 -0.79 Outer Logan Central -0.57 -0.58 -0.57 -0.56 Outer Goodna -0.70 -0.88 -0.73 -0.50 Outer Springwood -0.80 -1.04 -0.59 -0.79

The centres have been assigned one of five categories based on these intensification scores, a consideration of the centre’s characteristics of development, and their relative change in

28 As these scores are derived by averaging the z-scores of the various indicators, a score of 0 is approximately equal to the average of the composite scores.

164 comparison to the baseline changes. The results of this classification are described in Table 39 and discussed further below.

Table 39 - Classification of centre conformance

Category Centre Score based criteria Conforming Chermside Above average scores in all indicator groups. Indooroopilly Partially conforming – Cleveland Above average scores in both residential groups residential Toowong Toombul Partially conforming - Carindale Above average score in the employment group and average employment Upper Mount Gravatt scores in the residential groups Marginal conformance Capalaba Above average score in one group, but below average score Ipswich in another group Non-conforming Strathpine All scores average or below average Mitchelton Browns Plains Beenleigh Redcliffe Logan Hyperdome Wynnum Central Logan Central Goodna Springwood

Chermside clearly stands out with an overall score that exceeds other centres and a high intensification score across all factors. Indooroopilly also has above average scores in all categories, and both of these centres exceed baseline growth in all categories. These centres are therefore considered to be conformant with regional planning policy.

Cleveland, Toowong, and Toombul displayed above average conformance for residential aspects. Cleveland was the only outer centre to show a high degree of intensification and to exceed outer baseline change measures in all categories. Its employment intensification relative to other centres however, was less impressive. Toowong presented similar results and exceeded all baseline measures for middle ring centres, but this growth was similar to the levels of growth seen in other inner city areas. Toowong was the highest ranked centre in terms of compactness in 1996, and it remained so in 2016. It is therefore considered a conforming centre, but primarily in terms of its residential growth. Toombul displays very high levels of residential intensification, however it’s overall score is reduced by low rates of employment change. This is an interesting case, as although the actual quantum of employment change was small, the key changes to employment occurred in the mixed-use category rather than through the intensification of big box shopping centres and institutions

165 as was typical in centres with high employment growth. This is the result of converting a number of existing employment based uses into high-rise mixed use developments. Due to its intensive residential growth and high overall compactness rating, it is considered to be mostly conforming to regional policy intent despite its lower than average employment score.

Carindale ranks second on overall intensification, however this is primarily the result of its concentrated employment growth through the redevelopment of Westfield Carindale (see section 5.3). Its performance in residential intensification is less impressive, and although it added some additional low-medium and high density dwellings, its residential areas remained as primarily low density. For this reason, it has been classified as having partial conformance. This is a similar pattern for Upper Mount Gravatt. In both of these centres, residential growth and changes in density are similar or lower to overall baseline growth for middle ring areas.

Capalaba shows reasonable scores for intensification in residential uses however as discussed in section 5.1.1, its density changes are heavily influenced by the conversion of a number of large, low density allotments to low-medium density dwellings. Population and dwelling growth, whilst higher than most outer centres, was still only in keeping with baseline growth, and employment growth compared poorly with outer ring employment changes. For these reasons, Capalaba is considered to be a marginally conforming centre. Ipswich presents the opposite situation. Its employment changes are in keeping with activity centre policy, and the city centre saw several new or expanded employment based developments. However, the centre showed effectively no residential growth, and on this basis it was considered to be only marginally conforming.

The remaining centres exhibit a number of differences between intensification types. For most of these centres, the overall issue is one of a lack significant changes to the built form. For example, where higher density dwelling types are increased (and as previously, discussed the centres typically provide these dwellings more readily than surrounding non-centre areas), the quantity of the additional dwellings is relatively small. Wynnum and Mitchelton are somewhat different in that they added larger numbers of new dwellings, however almost half of these dwellings were of a low density form, which reduced their overall compactness scores. Browns Plains had very poor residential growth but did add a large amount of new employment based development. These new employment uses however involved large format uses with low employment rates, which did not score highly in terms of centre compactness measures.

Figure 22 plots the classification of centres spatially. As with the individual indicators, the conforming centres are mostly clustered within middle rings locations.

166

Figure 22 - Centre locations with conformance classification

167

The above discussion illustrates the nuances in how the different centres have developed over the past two decades. Categorising the centres into different types of conformance captures the key centre differences, however the creation of such categories inevitably requires a line between categories to be drawn somewhere. Although examples of some positive change can be observed in all the centres, regional centre policy intends for these centres to change at a scale that would make a material difference to the overall sustainability of the urban form. The ultimate line for conformance therefore requires positioning to reflect centres that have seen an appreciatively significant change to the urban form. The centres categorised as “non- conforming” are those that do not meet this threshold when considered holistically even though they may include some aspects of development that align with policy intent.

To further highlight the differences between the various types of centre conformance, the following sections provide a more detailed description of a selection of centres. Chermside was selected to demonstrate a conforming centre. Carindale is chosen to highlight the less ideal conditions inherent to the partially conforming category, as is Ipswich as one of the marginally conforming centres. Finally, Springwood is selected to show the characteristics of a centre with a low intensification ranking. These centres have all been designated as higher order, “principal” activity centres under the various iterations of the South East Queensland Regional Plan. Ipswich was also a key metropolitan centre under the RFGM plans, and Chermside, Carindale, and Springwood were all candidates for “major” centres under the RFGM.

5.5.1. Chermside

Key details Centre Type: Principal Regional Activity Centre Local Govt. Area: Brisbane City Overall compactness rank: 2nd (1.10) Intensification rank: 1st (1.53)

2016 1996 Population 7673 4333 Dwellings 4198 2345 Net residential density 37du/ha 22du/ha Dwelling IQV (0-1) 0.96 0.76 Estimated employment 13,809 7,958 Net job density 146 jobs/ha 85 jobs/ha Land use variation (0-1) 0.73 0.75 Active frontages 1,076m 633m

168

Figure 23 - Chermside 1997 (left) and 2016 (right) – (The State of Queensland, 2018), (NearMap 2016)

Chermside’s growth has significantly exceeded the other centres, and the extent of this intensification is clearly apparent in a comparison of aerial images from 1997 and 2016 (Figure 23). Chermside’s residential development has mostly come through the replacement of low density dwellings, although the Wesley Mission complex also underwent considerable redevelopment through the construction of multi-storied retirement housing. Playfield Street (Figure 24), immediately adjoining the east of Westfield Chermside, is becoming a concentrated strip of high density residential uses and mixed-use development. Other high density developments are also scattered throughout Chermside’s residential areas, and more than 20 additional sites were under development as of 2016. Large parts of the centre were also redeveloped into low- medium density developments. The most significant changes in employment were primarily the result of intensifications of existing uses, such as the upgrade to Westfield Chermside, as well as the Prince Charles Hospital. Shopping centre upgrades have resulted in floor areas more than tripling in area. Other employment changes are relatively small in comparison and include addition office and retail as part of a mixed- use office development (Figure 25), and expansions to health care services. As most of Chermside’s changes have been the result of the intensification of existing uses, mixed- use indicators show little change between the two time periods.

Figure 26 shows the locations of changed land uses (excluding sites under development in 2016) and the locations of points of interest such as sites that underwent intensive development.

169

Figure 25 - A new mixed use office building in Chermside (Google Street View 2017)

Figure 24 - High density dwellings in Chermside (Google Street View 2016)

170

Figure 26 - Chermside - changed uses and points of interest

171

5.5.1. Carindale

Key details Centre Type: Principal Regional Activity Centre Local Govt. Area: Brisbane City Overall compactness rank: 9th (0.80) Intensification rank: 2nd (0.73)

2016 1996 Population 4,186 3,782 Dwellings 1,754 1,395 Net residential density 17du/ha 14du/ha Dwelling IQV (0-1) 0.62 0.35 Estimated employment 4,941 1,997 Net job density 167 jobs/ha 74 jobs/ha Land use variation (0-1) 0.47 0.48 Active frontages 89m 72m

Figure 27 - Carindale 1997 (left) and 2016 (right) – source (The State of Queensland, 2018), (NearMap 2016)

Carindale’s intensification has been focussed entirely in and around the Westfield Carindale shopping centre. In terms of residential development, the centre saw the redevelopment of a temporary car parking area into a combined high density tower and townhouse complex (Figure 28). The only other residential development of significance was a new townhouse development to the west of the centre. The remaining residential area consists of low density dwellings on typically 600m2 lots. In terms of employment, 80% of the growth occurred through the redevelopment of Westfield Carindale. The expansion of the Belmont Private Hospital and a new strip to the north west of the shopping centre also added some employment. These areas are shown on Figure 29. Comparing this map with the map of Chermside’s land use change (Figure 26), shows the fundamental difference in how the two centres have developed over time.

172

Chermside displays development of a range of uses across the entire centre extent, while Carindale’s development is entirely dominated by a single big box shopping centre. Although Carindale has seen some higher density residential development and impressive employment growth in a concentrated space, the result is not entirely in keeping with the compact activity centre concept of “mini-CBDs”. It ultimately displays a poor variance and mix in uses and was therefore classified as being partially conformant.

Figure 28 - Apartments next to Westfield Carindale (left), and an example of the more typical low density uses (right) (Google Street View 2016)

173

Figure 29 - Carindale - changed uses and points of interest

174

5.5.1. Ipswich

Key details Centre Type: Principal Regional Activity Centre Local Govt. Area: Ipswich City Overall compactness rank: 5th (0.35) Intensification rank: 9th (-0.07)

2016 1996 Population 1,904 2,035 Dwellings 1,006 1016 Net residential density 17du/ha 15du/ha Dwelling IQV (0-1) 0.63 0.5 Estimated employment 12,719 8,156 Net job density 122 jobs/ha 80 jobs/ha Land use variation (0-1) 0.9 0.81 Active frontages 3,995m 3,700m

Figure 30 - Ipswich 1997 (left) and 2016 (right) – source (The State of Queensland, 2018), (NearMap 2016)

Ipswich is an historic town that was established during the colonial era. It features the highest mixed use rating of any centre, and has a full range of services from retail, to government administration. The town centre itself has a traditional main street character. Due to these aspects, Ipswich continues to rank highly as a compact activity centre. Intensification in the centre however has been almost entirely non-residential. Several commercial and community uses have been developed in Ipswich over the past twenty years. This includes the new Riverlink Shopping Centre (a big box style retail facility), a bulky good retail warehouse on the centre fringe, and a number of new office developments/conversions. The council owned Ipswich City Properties, arranged the development a new high rise office tower in the town centre, which was subsequently leased by the state government (Figure 31 and discussed further in section 8.2.2). Ipswich

175

Figure 31 - Ipswich's new office tower and traditional main street has also seen a significant hospital upgrade, and several new health care land uses have established in its proximity. Other institutional changes include a new court house, and school expansions. These land use changes have resulted a substantial increase in the number of estimated jobs. Residential growth however has been poor, and it is estimated that the centre has lost dwellings during the past two decades. This been the result of residences being replaced for commercial development or being converted into other uses (Figure 32). Although Ipswich has seen some higher density residential uses emerge, the number of

Figure 32 - House to office conversions (left) and the Oaks Aspire residential tower (right) (Google Street View 2017)

dwellings created has been less than the low density dwellings lost due to redevelopment or other reasons. Notable among Ipswich’s residential development is the Oak’s Aspire high rise tower (Figure 32). Most of the units within the tower however are being used for short term accommodation, and the ABS only recorded 27 occupied high rise

176 dwellings in the 2016 census. Higher density residential growth in activity centres has been a key component of the policy since its inception. Ultimately, it was this lack of residential growth that resulted in Ipswich being considered as marginally conforming to centre policy, even though it added some new higher density dwellings. Figure 33 shows several new commercial developments across the centre, but a dearth of new residential uses.

177

Figure 33 - Ipswich - changed uses and points of interest

178

5.5.1. Springwood

Key details Centre Type: Principal Regional Activity Centre Local Govt. Area: Logan City Overall compactness 17th (-0.32) rank: Intensification rank: 19th (-0.80)

2016 1996 Population 2,714 2,874 Dwellings 1,101 1,148 Net residential density 17du/ha 17du/ha Dwelling IQV (0-1) 0.59 0.6 Estimated employment 5,619 4,942 Net job density 69 jobs/ha 62 jobs/ha Land use variation (0- 0.76 0.72 1) Active frontages 372m 211m

Figure 34 - Springwood 1996 (left) and 2016 (right) – source (The State of Queensland, 2018), (NearMap 2016)

Springwood was the most poorly ranked centre in terms of intensification and as might be expected, there have been few changes to the centre over time. The centre is bisected by the Pacific Motorway. To the west it is primarily comprised of light industrial uses, and bulky goods retail. There has been some intensification of these uses, however these changes are marginal. The eastern side is made up of offices, big box shopping centres, and includes bulky goods retail. Most of the residential uses are also on the east. Although two small scale medium density residential developments were observed, these uses were completed in mid-2016 and as the census did not include any three storey

179

apartments, this suggests that the dwellings were not occupied at the time. Apart from these, no other dwelling intensification was observed. The most significant changes to the centre have been the construction of two, multistorey hotels (Figure 35) which represent the centre’s largest employment growth. Springwood has also seen some small scale office development, and minor upgrades and extensions to some of the existing shopping centres. Overall however, the changes to the centre are comparatively minor and this sparse level of development can be observed on the map in Figure 36.

Figure 35 - Springwood Towers Hotel behind a new mixed-use office (left), and a new small scale office on Cinderella Drive (right) (Google Street View 2018)

180

Figure 36 - Springwood - changed uses and points of interest

181

5.6. Conclusion

The previously highlighted examples demonstrate that there are not only fundamental differences between how the centres have changed over time, but also in terms of their core physical structures and functions. These factors shape potential development possibilities and the end result is a complex picture of change that includes wide variations in patterns of development. This part of the research has evaluated whether these observed changes conform to metropolitan scale policy to create a network of compact centres, characterised by higher density living and mixed clusters of uses providing employment and services. Conformance in this case is considered in positivist terms of progress; that plans are intended to be effectuating devices and after two decades of continuous policy to create more compact activity centres, planned intentions should be evident in physical changes to the built form. When investigated on this basis, the designated centres showed poor results with only two of the nineteen centres displaying substantive progress in becoming more compact in all nominated categories. Five other centres also showed promising partial conformance. Centres in outer ring locations typically conformed more poorly compared middle ring centres. Here, the lack of conformance was primarily the result of limited change rather than of changes that were directly contrary to compact activity centre policy, suggesting metropolitan policy has an implementation problem.

The principal form of implementation proposed by the plans is to adjust land use regulatory provisions to enable development that conforms to plan intentions (see section 2.2). Having demonstrated a lack of conformance with overall objectives for compact centres, Chapter 6 expands the research to consider implementation mechanisms more closely, and investigates the degree to which regulations have changed in line with metropolitan policy intentions. This provides an indication as to the performance of metropolitan policy, and also establishes a basis from which relationships between plan conformance and policy change can be investigated. These relationships, as well as links to other factors that may explain the lack of implementation, are considered in Chapter 7.

The results also highlight some of the complexities involved in making reliable comparisons of urban intensification. The different nature of the studied cases, and sometimes even the nature of a particular development or site, led to situations where the results from the indicators required further qualification. The assembled indicators would therefore benefit from further analysis and refinement, particularly in terms of the development of methods to evaluate centre policy. By including additional cases, the selected measures could be subjected to more statistical testing, as well as permitting the development of more complex models that can explain their results. The development of suitable indicators also proved challenging, with

182 new methods being required in order to overcome data limitations (section 4.2.2). Future research would therefore benefit from the expansion of existing government open data policies, especially in terms of detailed land use data and historical data sets. These aspects, and other recommendations from the research are discussed further in Chapter 8.

183

6. Performance of policy in influencing the development and application of land use regulations

Compact activity centre policy is intended to primarily be implemented by changing land use regulations to be more supportive of uses that contribute to the policy’s objectives, such as higher density residences and more intensive commercial uses. These regulations are then applied to development proposals to change uses. The previous chapter demonstrated that after twenty years of consistent policy, most activity centres conformed poorly to policy intent. The research now seeks to examine possible reasons for this result. An important area of inquiry of this nature must therefore be to investigate the application of the key implementation mechanism. To do so, this chapter addresses the research questions of how metropolitan planning policy for compact activity centres has performed in influencing land use regulations, and how actual land use changes conform to these regulations. As previously discussed in section 3, evaluating plan performance involves determining whether a plan or policy was used when making decisions relevant to the plan. Rather than directly considering such decisions, the research seeks evidence of decisions in changes that can be observed in land use regulations. Historical land use regulatory documents and maps are sourced, and their content is firstly examined for references to regional activity centre policy. Next, the regulatory intent of these documents are investigated and assigned “development intensity scores” (DIS) for a variety of land use types in order to classify the intensity of development permitted by the regulations. The DIS are applied to all properties in the study area in order to quantify the extent of changes to the intended intensity of development, and whether these changes accord with compact activity centre principles. Finally, the previously developed land use database is used to compare actual land uses and land use change to the type and intensity of use intended by planning regulations.

The results demonstrate that compact activity centre policy is performing strongly. Over time, regulations for all of the centres changed to include direct references to activity centre policy, and the regulations themselves changed to be more permissive of uses aligned with compact activity centre principles. This indicates that the policy has been well integrated in local government land use regulations both at the strategic and code level. Where land uses changed, new uses are highly likely to be in accordance with the land use regulations demonstrating that local governments are applying land use regulations as intended by making decisions on individual developments that reflect regulatory intent. Despite this, only a small proportion of sites intended for further development actually changed. This reinforces the results of Chapter 5 which shows poor conformance is mostly due to a lack of change. The poor conformance of actual land use change, combined with the strong performance of policy

184 in influencing land use regulations, highlights that the key implementation mechanism for compact activity centres (the regulation of land use) is insufficient to effect the desired change to the built form. The following sections present the detail of these results, which subsequently inform further analysis of factors that explain centre intensification in Chapter 7.

6.1. Content analysis of the performance of local government land use regulations

This section reports on the results of a basic content analysis of local government land use regulations. The regulations in place for each activity centre in approximately 1996, 2006, and 2016 were examined to determine if they reflected regional policies for compact activity centres. The purpose of this is to evaluate the performance (i.e. whether regional policy was used in decision making in plan formulation) of regional policy. Where the regulations directly reference regional policy and/or use regional policy terminology, the regional policy is considered to be performing, as there is evidence it was used in the formulation of the regulations.

References to regional policies for activity centres were assigned one of four categories:

1. No reference to regional policy found in the document

2. Reference found to regional policy, but not in relation to activity centres

3. Regional policy for activity centres is described at a strategic level, however the centre is not specifically nominated

4. Regional policy for activity centres is specifically discussed and the centre is specifically nominated

The results of this categorisation are shown in Figure 37.

185

Figure 37 - Planning regulations and their type of reference to regional policy, 1996, 2006, and 2016

3. 2. 4. 1. Strategic centre SEQRP reference, Centres No SEQRP reference, but but not activity specificaly reference centre not centres nominated nominated

Albert Shire Council – Beenleigh; Brisbane City Council Logan City Council - – Carindale, Redlands City Browns Plains, Logan Chermside, Council – Capalaba, Central, Logan Indooroopilly, Cleveland; 1996 Hyperdome, None Mitchelton, Toombul, Ipswich City Council Springwood; Toowong, Upper Mt. – Ipswich, Springfield; Pine Rivers Council – Gravatt, Wynnum; Pine Rivers Council – Strathpine; Ipswich City Council - North Lakes Redcliffe City Council Goodna - Redcliffe

Ipswich City Council - Ipswich, Spingfield; Gold Coast City Brisbane City Council Council – Beenleigh; – Carindale, Logan City Council – Chermside, Browns Plains, Logan Indooroopilly, Central, Logan Redcliffe City Council None Mitchelton, Toombul, Hyperdome, 2006 - Redcliffe Toowong, Upper Mt. Springwood; Gravatt, Wynnum; Pine Rivers Council – Ipswich City Council - North Lakes, Goodna Strathpine; Redlands City Council – Capalaba, Cleveland

2016 None None None All centres

Based on the 1996 era documents, most local governments were quick to include aspects from the RFGM process into their planning regulations, with regulations covering 14 of the 21 centres including a reference to the regional policy. Only five of the centres however were specifically mentioned as being centres as defined in the RFGM. Regardless of regional policy however, all local governments were undertaking some form of centre based planning for these locations and made use of a hierarchy of centre types. Consistent with the narrative of

186 the evolution of compact activity centre policy (section 2.1.1), this planning was usually based around planning for commercial and industrial uses, unless the centre was in the fourth category. By 2006, all of the planning regulations included references to regional planning. The Redcliffe City Planning Scheme 2005 is unique as the only planning scheme that makes reference to regional planning policies but not the activity centre policy, however by 2008 this planning scheme was amended to include a reference to the area being a major activity centre under the SEQRP. At this time, planning for Brisbane’s centres continued to reference the RFGM policies at a strategic level. With the exception of Wynnum, the Brisbane City Plan 2000 (as amended in July 2006) had centre specific local plans covering all of the nominated centres. Although the local plans did not use RFGM terminology, they all featured compact activity centre based principles for future development, and were defined as “major centres” in the city’s strategic plan. Planning for Wynnum however was more residential than centre focussed. This reflects Wynnum’s lower role in the Brisbane centre hierarchy of the time. Goodna presented a similar situation. The 1996 and 2006 era Ipswich City Council planning schemes made specific mention of the RFGM however Goodna’s role as a major centre was not specifically discussed. The council did apply compact activity centre policies to the Goodna centre, however these appear to be more generalised rather than inspired directly by the regional policy. The 1996 era Ipswich City Council planning regulations note that Redbank was preferred as the predominate centre of eastern Ipswich. By 2016, all local governments specifically referenced each centre in terms of their role in the regional plan.

Local government land use plans are the principal mechanism for implementing compact activity centre policies. The results from this analysis show that although the timing varied between different local governments, the regional policy performed well in terms of aligning land use plans with regional policy intent. The direct inclusion of terminology, principles, and references to the regional plans themselves indicate that the policy “messages” were received by local governments and, that over time, decisions had been made to include these messages in their own plans.

One aspect to consider is the potential issue of coercion. From the release of the first SEQRP in 2005, local governments were required to incorporate regional planning objectives when creating or amending their planning schemes. The early adoption of the regional centre policy concepts and terminology shows that Redlands, Ipswich, Pine Rivers, and Brisbane councils were referencing regional policy in their land use planning prior to this date, and were therefore undertaking these changes voluntarily. Aside from Redlands and Brisbane, these councils were applying the policy differently across their centres, suggesting that the policy was known, and that councils were choosing to selectively apply it; a performance success.

187

Gold Coast City Council also voluntarily adopted the regional plan by specifically designating Beenleigh as a regional activity centre in their 2003 planning scheme. The case for voluntary adoption in Logan and Redcliffe councils is less clear as these local governments did not include the regional activity centre policies in their planning documents until after the release of the SEQRP.

This analysis has shown evidence that all local governments in the study area eventually referenced regional activity centre policy in their planning regulations. This suggests that the regional policy was used in decision making when creating land use regulations as intended by the regional plans, and that the regional policy is therefore performing well. However, this content analysis only captures the wording used by planning regulations at a strategic level. Local government planning schemes typically regulate development through the application of zones and local plans on individual properties that are supposed to reflect this strategic intent. By quantifying the intensity of development permitted by these mechanisms at different times, it is possible to determine if the performance of regional policies for activity centres also extended to significant changes in actual rights in land. The results of these aspects are discussed in the following section

6.2. Changes to local government land use regulations and their alignment to activity centre policy

This section discusses the results and analysis of changes to local government land use regulations, and the degree to which these changes align with compact activity centre policy. Changing local government land use regulations to be permissive of the type and scale of land uses envisioned for compact activity centres, is a key implementation mechanism for regional level policy (section 2.2). The previous section demonstrated that regional level policy performed well in terms of local governments recognising compact activity centre policy at a strategic level in their land use regulations. The overall purpose of these regulations however, is to regulate development on individual development sites. To do so, Queensland land use regulations make use of zones and/or other spatially defined area classifications to apply strategic policy intent to individual properties, thereby limiting the type and scale of particular land uses. As described in section 4.2.3, Development Intensity Scores (DIS) were developed to quantify the intensity and type of land use permitted by land use regulations. A description of the DIS definitions is provided in Table 10, p110. By using a DIS, it is possible to compare regulations over time and across different jurisdictions, and whether these regulations are changing in line with compact activity centre policy.

188

A review of regional policy (section 2.2) revealed that planning regulations that reflect the regional policy for compact activity centres should show evidence of:

1. Greater proportions of land allocated for higher density residential uses;

2. Greater proportions of land allocated for more intensive commercial uses, including a provision for bulky goods retailing;

3. Reduction in the proportion of land for larger scale industrial uses; and

4. Greater proportions of land that permit both residential and commercial development (i.e. mixed use development)

A DIS was determined for every property in the study area (n=44,063) based on an analysis of the various planning schemes current at approximately 1996, 2006, and 2016. The DIS measures the potential intensity of residential, commercial, industrial, and bulky goods retail uses on each property. By summing the land areas associated with each DIS score at each point in time, it is possible to assess whether the regulations have changed to be more permissive of the land uses required to achieve compact activity centres. These changes are summarised in Table 40.

Centres in which regulations that changed in-line with activity centre policy intent are considered to be positively performing. Centres were considered to have marginal positive performance when only a minor change to regulations was observed (change of <5% of centre area), or where the regulatory changes did not show a clear picture of overall intensification/reduction (e.g. Capalaba’s commercial regulations which showed a mix of increases and decreases across different commercial scores). Centres that saw regulatory changes contrary to activity centre policy were considered as policy exceptions.

This table shows that in most instances, land use regulations are changing to match regional policy intent. 11 of the 19 centres showed no exceptions to the expected policy changes, and only three centres had more than a single exception. Positive performance is particularly evident in the residential category, where all centres saw changes that permitted higher density residential development. Changes to regulations for commercial development also reflected positive performance with 15 of the 19 centres recording positive changes, and the remaining four centres showing marginal positive change. A greater amount of marginal results are recorded for the bulky goods retail and industrial categories, with generally smaller changes in area being observed. These changes however are typically positive, with only four exceptions recorded in each category. The more marginal positive performance in these categories may be related to the reduced prominence given to these aspects of compact city

189 development in the regional policy. More likely however is that these aspects reflect specialised uses associated with different centres types, with outer centres typically having more land available for large format uses such as bulky goods retail and industry. The mixed use category also typically shows marginal changes, however this reflects that mixed use development was generally already well accommodated by land use regulations in 1996, and there was therefore little capacity for further change (section 6.2.5).

Table 40 - Summary of changes to land use regulations by centre, 1996 to 2016

Location Centre Name

Residential Commercial good Bulky retail Industrial use Mixed

Inner Toowong N/A Carindale N/A Chermside Indooroopilly Middle Mitchelton Toombul Upper Mount Gravatt Wynnum Central Beenleigh N/A Browns Plains Capalaba Cleveland Goodna Outer Ipswich Logan Central N/A Logan Hyperdome Redcliffe Springwood Strathpine

Legend positively performing policy exceptions marginally performing

N/A represents centres that had no land (or very little - <1% of area) regulated for industrial development in 1996, and could therefore not reduce the amount of industrial land.

The results for each category are described and discussed in further detail below.

190

6.2.1. Residential regulatory changes The residential category shows particularly strong positive performance with regional policy intent, with all centres showing regulatory change to be more permissive of higher density residential development. Figure 38 and Figure 39 shows the differences in the DIS for residential uses between 1996 and 2016 for each location and centre in the study area. These differences are expressed in terms of their percentage of the total land area for each centre. As described in Table 10 (p110), the residential DIS is a scale of the intensity of permitted residential use ranging from 1 (no residential development permitted), through to 7 (apartment buildings seven stories and above).

Figure 38 - Differences in land area of residential DIS between 1996 and 2016, as a percentage of total land area, by location

Inner

Middle Location

Outer

-30.0% -20.0% -10.0% 0.0% 10.0% 20.0% 30.0% Percentage change Res Score 1 Res Score 2 Res Score 3 Res Score 4 Res Score 5 Res Score 6 Res Score 7

Theses chart shows a clear picture of land use regulations changing to be more permissive of higher density residential uses, particularly those in DIS 6 and 7 (which permit 4-6 storey and 7 plus storey apartment development respectively). The results (section 10.5.1) also show that over time, the degree of permissiveness has increased, where changes between 1996 and 2006 were more likely to be in the form of increases to DIS 5 and 6, and changes between 2006 and 2016 saw a larger increase in DIS 7. The changes show a consistent increase of permitted residential intensity across all centres, although the extent of change does vary. Overall, the

29 Residential DIS 2 and 3 represent low density residential uses typically made up of detached dwellings. Residential DIS 4 and above represent attached housing. To more easily differential between detached and attached types of residential uses, these classifications are distinguished by the green/red colouring schemes in the chart.

191 outer centres show greater proportional increases in land zoned for higher densities compared to middle ring centres. Browns Plains is unusual due to the large proportion of land which changed to restrict residential development. This change was primarily the result of the development of a new large scale industrial estate that accounts for approximately 25% of the centre’s land area, and in which residential uses were limited (section 5.4.1). If this area is excluded, Browns Plains shows regulatory changes that are also primarily more permissive of higher density residential uses.

Figure 39 - Differences in land area of residential DIS between 1996 and 2016, as a percentage of total land area, by centre

Although the overall trend of more intensified residential land use regulations is consistent, key differences emerge in term of regulations that saw a reduction in their extents. Changing

192

existing low density areas to higher density is a core principal of the compact city concept and its response to urban sprawl. To align with this, reductions in the amount of land zoned for low density development would also be expected. However, examining the areas that changed to permit higher density development highlights some interesting differences between the approaches used by different local governments. Figure 38 shows that while outer areas typically converted restricted or low density residential areas to higher densities, changes in the inner and middle centres of Brisbane City Council predominantly came from areas that were already zoned for attached housing.

Brisbane City Council is rezoning properties for residential development differently to other local governments. This council is typically selecting areas for higher densities from locations that are already zoned for attached housing, albeit at medium densities. One explanation might be that Brisbane’s centres are already more developed and therefore have fewer existing areas zoned for low density from which they can change. However, this is not evident in the data, which shows that in 1996 middle and outer centres had similar proportions of land with residential DIS 3 and DIS 4. Recent research from Los Angeles shows that residential “upzoning” was most likely to occur in places with lower “political resistance” (locations with lower proportions of homeownership), with fewer valuable amenities, and lower intensities of existing zoning (Gabbe, 2018). The research concluded that this pattern represents existing properties owners attempting to prevent changes to their neighbourhood by organising to prevent rezonings for higher density development. Australian research also suggests that community opposition presents a significant barrier to higher density development (Cook et al., 2012; Nematollahi et al., 2016; Ruming, 2014). One possible explanation is that the ownership characteristics of Brisbane’s existing low density areas differ from the outer centres, and rezonings are more likely to be subject to resident opposition and therefore avoided in favour of areas where expectations for higher density have already been established (such as the DIS 4 areas). Such a phenomenon could seriously limit any future rezoning possibilities in centres where all the “lower hanging fruit” has already been rezoned (e.g. Carindale), and therefore compromise rezoning as a method to implement compact activity centre policy. This is therefore a matter worthy of further research, although it is beyond the scope of this dissertation.

6.2.2. Commercial regulatory changes Changes in the commercial category also showed mostly positive performance, where there was an overall increase in the intensity of permitted commercial development, and only four of the outer centres with marginal changes. Figure 40 shows the changes to the commercial

193

DIS as a percentage of the total land area of inner, middle and outer ring centres30. The most significant changes in all locations were increases in DIS 6 which indicates allowances for large scale retail and office complexes in structures five storeys and above. Overall the change to commercial regulations affected a relatively small proportion of each centre’s area compared to the residential changes (12% average change vs. 27%). The general trend across centres was to rezone existing commercial uses to more intensive uses rather than create new commercial sites, with only five of the centres showing greater than 5% of their areas changing from no or limited commercial uses. Brisbane council’s inner and middle centres drew mostly from sites already zoned for intensive commercial use, to permit increased areas for high-rise (5 storey and above) commercial development. These centres also reduced limited commercial development opportunity (COM 1 and 2) areas to a greater extent than outer areas. There was a greater variation in commercial regulatory changes in outer centres (Figure 41), however the overall trend was for outer centres to convert existing small scale centre zoned areas (DIS 3) to permit more intensive commercial uses. This difference between locations however reflects the amount of land in the DIS 3 category in 1996, with outer centres having a far higher proportion of DIS 3 land compared to middle centres (12% of centre area vs. 1%).

Figure 40 - Differences in land area of commercial DIS between 1996 and 2016, as a percentage of total land area, by location

Inner

Middle Location

Outer

-10.0% -8.0% -6.0% -4.0% -2.0% 0.0% 2.0% 4.0% 6.0% 8.0% 10.0% Percentage change

COM Score 1 or 2 COM Score 3 COM Score 4 COM Score 5 COM Score 6

30 DIS 2 for commercial uses represents very small scale, individual commercial uses such as general stores. This score is commonly applicable to a large proportion of the centre areas and although permissive of some commercial uses, they are highly limited by restrictions on scale and the need to protect residential amenity. These scores have been combined with DIS 1 to enable easier comparisons between more intensive commercial zonings (DIS 3 to 6). Full DIS results by centre, including DIS 2, can be viewed in section 10.5.

194

Figure 41 - Differences in land area of commercial DIS between 1996 and 2016, as a percentage of total land area, by centre

6.2.3. Bulky goods retailing regulations Changes to bulky goods retailing regulations reveal some centres have changed contrary to regional policy, although these instances all involved less than 5% of the centres’ areas. Most of the outer centres showed positive performance however outer centres also had significantly greater proportions of regulations supportive of bulky goods retail in 1996, compared to middle and inner centres (21% of centre area vs. 7%). The various iterations of the regional plans have considered bulky good retailing as a special category of commercial use. The regional plans acknowledge that these types of uses require large areas of land and they can

195 therefore be difficult to accommodate within activity centres. The planning policy for bulky goods retail therefore states that these uses should ideally locate within centres, however this form of retail is considered to be an acceptable “out of centre” use. This was reflected in the 1996 regulations where large areas31 for bulky goods retailing were found only in outer locations. Some of these outer centres however included significant areas (>30% of total centre area) supportive of bulky goods retailing such as Browns Plains, Capalaba, Springwood, and Strathpine. The majority of centres saw less than 5% of their land area change in relation to requirements for bulky goods retail. Figure 42 shows the changes to the bulky goods retail DIS as a percentage of the total land area of inner, middle and outer ring centres. The overall trend shows a slight increase in the amount of land where bulky goods retailing is permitted. However, the extent of change varies considerably between centres. The majority of the middle centre change is the result of changes in Chermside and Upper Mount Gravatt where approximately 50ha in each centre was changed to permit limited scale bulky goods retailing (Figure 43). This change does not appear to have been targeting bulky goods retailing specifically however as it was the result of the land being added to the centre frame zone which allows a very broad spectrum of potential commercial uses, including limited bulky goods retailing.

Figure 42 - Differences in land area of bulky goods retail DIS between 1996 and 2016, as a percentage of total land area, by location

Inner

Middle Location

Outer

-10.0% -8.0% -6.0% -4.0% -2.0% 0.0% 2.0% 4.0% 6.0% 8.0% 10.0% Percentage change

BGR Score 1 BGR Score 2 BGR Score 3

31 (>10% of total centre area)

196

Figure 43 - Differences in land area of bulky goods retail DIS between 1996 and 2016, as a percentage of total land area, by centre

6.2.4. Industrial regulations Regional policy considers large format industrial uses and warehousing to be important employment generating uses, but not associated with the type of use expected in activity centres. Changes to land use regulations that align with this objective should therefore show reductions in the extent and/or intensity of permitted industrial uses. Figure 44 shows the changes to the industrial DIS as a percentage of the total land area of inner, middle and outer ring centres. This chart shows that all areas recorded reductions in the intensity and/or amount of land zoned for industrial uses. Looking at individual centres, all of the inner and

197 middle centres reduced industrial zonings, however the results were more mixed in the outer centres (Figure 45). Ipswich council intensified industrial land (<8% of centre area) in DIS 3 and 4 in both of its centres, while Redland council introduced some limited industrial uses in Cleveland while reducing the intensity of industrial areas in Capalaba. Logan council made only minor changes (<3% of centre area). The larger scale changes in Browns Plains occurred mostly in the Brisbane council controlled area in the northwest of the centre where a new industrial estate was established. This resulted in some large areas being formalised for industrial uses, while others were regulated for environmental conservation.

Figure 44 - Differences in land area of industrial DIS between 1996 and 2016, as a percentage of total land area, by location

Inner

Middle Location

Outer

-8.0% -6.0% -4.0% -2.0% 0.0% 2.0% 4.0% 6.0% 8.0% Percentage change

IND Score 1 or 2 IND Score 3 IND Score 4 IND Score 5

32 DIS 2 for industrial uses represents very small scale, individual uses such as repair shops, bakeries etc. This score is commonly applicable to a large proportion of the centre areas and although permissive of some very low impact industrial uses, the extent of these uses is highly limited by restrictions on scale and the need to protect residential amenity. These scores have been combined with DIS 1 to enable easier comparisons between more intensive commercial zonings (DIS 3 to 5). Full DIS results by centre, including DIS 2, can be viewed in section 10.5.

198

Figure 45 - Differences in land area of industrial DIS between 1996 and 2016, as a percentage of total land area, by centre

6.2.5. Mixed use regulations Areas of mixed use were determined by summing the area of properties in each centre which had a residential DIS 4 or greater (i.e. permissive of attached dwellings) and a commercial DIS 3 or greater (i.e. permissive of at least small shopping centres). All locations showed an increase in the amount of land that is permissive of mixed use development (Figure 46), with outer areas showing an overall greater increase, almost doubling the 1996 value. However, by considering individual centres (Figure 47), it can be seen that this increase is mostly due to very large increases in a few centres (Browns Plains, Logan Central, Logan Hyperdome, and Strathpine). The other outer centres show more marginal increases and even some reductions

199

(Cleveland and Ipswich). The majority of outer centres also had a larger proportion of mixed use areas compared to middle centres, in both 1996 and 2016.

The overall marginal changes to mixed use regulations is mostly explained by the nature of existing regulations. In 1996, 90% of land (on average) regulated for significant commercial uses in inner and middle centres already permitted the development of attached residential uses. In outer areas (excluding the three centres that did not permit mixed use development), 75% of commercially zoned land allowed mixed use development. By 2016, 96% of inner and middle commercial areas, and 93% of outer commercial areas, permitted mixed use development.

Figure 46 - Percentage of land permissive of mixed use development in 1996 and 2016, by location

25%

20%

15%

10%

5% Pecentage Pecentage of landarea mixed for use 0% Inner Middle Outer Location

1996 2016

200

Figure 47 - Percentage of land permissive of mixed use development in 1996 and 2016, by centre

45%

40%

35%

30%

25%

20%

15%

10%

Percentage Percentage of landtotalare for mixed use 0%

Ipswich

Goodna

Redcliffe

Toombul

Capalaba

Toowong

Carindale

Cleveland

Beenleigh

Strathpine

Springfield

Chermside

Mitchelton

NorthLakes

Springwood

Indooroopilly

LoganCentral

BrownsPlains

Wynnum Central Wynnum

LoganHyperdome Upper Mount Upper Gravatt Inner Middle Outer Centre and location

1996 2016

6.2.6. Summary In section 6.1 it was shown that all local governments have changed their land use regulatory documents to include strategic references to activity centre policy. In this section, the regulations were examined to confirm whether local governments used these strategic intents to inform material changes to development regulations. Based on the above analysis, it can be concluded that local governments are changing their land use regulations to align with activity centre policy. Although some patterns appear in terms of how regulations are changing between local governments (such as the differences between Brisbane’s changes to residential regulations compared to other local governments), it is also common to see local governments adjust regulations differently for different centres. Such variations are expected and consistent with policy performance provided there is evidence that the policy is being considered. The designated activity centres show significant differences in terms of the existing structures and land uses, and it is to be expected that different land use regulations will be devised to align with the characteristics of each centre. All centres however demonstrated positive performance in the key categories of residential and commercial intensification, with some centres showing exceptions only in less significant categories (bulky goods retail and industry). Overall, it is clear that local governments are making decisions on

201 land use regulations both in terms of strategic intent as well as in terms of the regulations themselves, which are generally well aligned with regional compact activity centre policy. This analysis is based on the outputs of regulatory decision making, rather than on the processes of decision making itself. As such, how regional policy was actually considered when making regulatory decisions is not known. Changes to the regulations themselves however, imply a strong degree of policy performance, with activity centre policy being well reflected in changes to regulations over time. Land use regulations are the key implementation mechanism for activity centre policy. Based on these changes, it can be concluded that the regional policy is performing well, with local governments using activity centre policy to inform regulatory change. The remaining question then is whether actual land uses and land use changes respond to these regulations. This matter is addressed in the following section.

6.3. Conformance of land use changes to planning regulations

The previous section demonstrated that land use regulations in activity centres have been changing over time to reflect regional level policy to create more compact centres. This section presents and discusses the results of an analysis of conformance of land use and land use regulations. Using the previously created Google Street View land use database, Development Intensity Scores, and building footprints, categories of conformance are assigned to every property parcel and site within the study area. Conformance is firstly considered across all areas within each centre at the beginning and end of the study period (1996 and 2016). This analysis reveals that land uses are typically highly conformant to land use regulations in both 1996 and 2016. Next, all sites where land use changed occurred were considered in terms of how this change conformed to the land use regulations relevant at the time of the change. Land use changes also proved to be highly conformant to land use regulations. This result implies that local governments are making decisions on development on the basis of their land use regulations and that these regulations are therefore performing well. Non-conformances primarily consisted of under-development, and few sites developed at a scale or type not intended by planning regulations. These results demonstrate that where development is occurring, land use regulations are working as intended by limiting uses to the desired type of use.

The GSV land use database and DIS scores were combined using GIS. Each land parcel in the study area was then classified according to its degree of conformance to land use regulations. Each parcel was classified as either:

202

• Conforming – these parcels are where the existing land use aligns with a DIS that reflects the use. For example, a commercial use such as strip mall or main street would be conforming if located on a parcel that had DIS that is permissive of commercial uses of this nature (i.e. equal to 3 or greater).

• Under developed – these parcels are where the existing land use has a scale or type that is less than that intended by the land use regulations. Examples include detached dwellings in high density residential areas, or vacant land that is permissive of other uses.

• Exceeding – these parcels are where the existing land use is of a type or scale that conflicts or exceeds the land use regulations. For example, an industrial use located on a site that has a DIS that does not permit industry (i.e. equal to 1), or apartments in a location intended for detached dwellings.

The classification was undertaken based on the estimated land uses in 1996 and 2016 using the land use regulations corresponding to these years33. The land area for each classification was then summed by each activity centre. Table 41 presents the results as a percentage of centre area for each conformance classification.

33 Note – as previously discussed, planning schemes dating to exactly 1996 could not always be obtained. The 1996 regulations used in this study represent regulations that were as close to 1996 as could be sourced. A list of the planning schemes used to determine the 1996 DIS are listed in the table in section 10.4.

203

Table 41 - Classification of land use conformance to land use regulations as a percentage of centre area, 1996 & 2016

1996 2016

Under Under developed Exceeding Conforming Under developed Exceeding Location Centre Name Conforming Inner Toowong 58.7% 37.4% 3.9% 71.0% 27.0% 2.0% Carindale 94.1% 5.5% 0.4% 96.9% 2.1% 1.0% Chermside 82.8% 15.6% 1.6% 82.6% 17.1% 0.3% Indooroopilly 80.8% 16.0% 3.1% 79.4% 19.3% 1.3% Middle Mitchelton 89.4% 9.5% 1.1% 84.2% 11.0% 4.9% Toombul 64.1% 34.6% 1.3% 62.5% 35.4% 2.1% Upper Mount Gravatt 82.9% 16.3% 0.8% 75.9% 22.0% 2.1% Wynnum Central 85.6% 13.5% 0.9% 82.1% 15.3% 2.6% Beenleigh 64.1% 33.8% 2.1% 58.2% 39.3% 2.5% Browns Plains 49.7% 50.2% 0.1% 87.8% 9.9% 2.3% Capalaba 79.1% 13.2% 7.7% 71.5% 18.1% 10.4% Cleveland 62.8% 33.3% 3.9% 75.5% 19.3% 5.2% Goodna 66.2% 24.1% 9.7% 71.5% 20.1% 8.4% Outer Ipswich 62.9% 30.9% 6.1% 63.2% 26.1% 10.7% Logan Central 91.5% 6.1% 2.4% 90.2% 9.6% 0.2% Logan Hyperdome 87.7% 9.6% 2.7% 87.1% 12.8% 0.1% Redcliffe 89.3% 7.8% 2.9% 80.0% 17.4% 2.5% Springwood 94.9% 4.8% 0.3% 70.0% 28.4% 1.7% Strathpine 59.5% 24.0% 16.5% 73.6% 24.1% 2.3%

Although there are some large variations between some centres, in both 1996 and 2016 the overall level of land use conformance is high (mean 73%). Of the non-conforming areas, land uses that exceeded the regulations typically made up a very low proportion of each centre (mean 4%). Changes to these proportions over time were also minimal, and only four of the centres changed by more than 10% in any category of conformance. Browns Plains and Springwood saw the largest changes over the two decade study period, and are illustrative of the different ways in which conformance can change over time. Figure 48 shows the categories of land use conformance in Browns Plains in 1996 and 2016 respectively. In Browns Plains, large areas of under-developed land were mostly vacant in 1996. These areas represent sites that had yet to realise the development potential accorded by the regulations. By 2016, much of this area had been developed in line with land use regulations or the regulations had changed to reflect the new uses, making these areas predominately conformant. Regulations had also changed in new locations to support more intensive uses, changing the category of conformance in these locations to “under developed”. Conformance change in Springwood

204 were almost entirely of this latter form. As discussed in section 5.5.1 Springwood underwent very little change during the study period. The large increase in “underdeveloped” land is therefore the result of land use planning regulations changing over time to be permissive of more intensive development (Figure 49).

205

Figure 48 - Browns Plains land use conformance, 1996 and 2016

206

Figure 49 - Springwood land use conformance, 1996 and 2016

207

This interplay between land use and regulatory change highlights some of the difficulties associated with considering plan conformance in this manner. In Queensland, land use regulations are aspirational in that they are intended achieve desired future land uses that align with strategic intent. Although locations where land use and regulation align can be said to be conforming, non-conformance is not necessarily an issue that is indicative of planning failure. Loh (2011) notes that in many instances non-conformance is a natural stage of the property development cycle that simply represents uses that have yet to change in-line with the planned intention. Considering conformance based on the use and regulations present at the same point in time, although helpful to quantify development potential, does not adequately account for this problem of temporal alignment. It is therefore necessary to consider conformance in terms of the situations where development resulted in a land use change.

The previously developed land use database records use changes for every property parcel in the study area. By combining this database with the previously developed building footprints it is possible to analyse conformance in terms of the number of changed development sites, rather than land area. As discussed in section 4.2.3, such an analysis better reflects land use planning regulatory decisions, and the proportion of which are conformant. Each site of land use change can then be compared with the corresponding DIS from the land use regulations that were likely to have been applicable at the time of the use change. For example, a use that changed in 2010 is assessed for conformance with the 2006 DIS, while a use that changed in 2000 would be assessed against the 1996 DIS. Such an approach resolves the temporal alignment issue in terms of both use and regulatory change and reflects how land use regulatory decision making occurs in the Queensland planning system.

The following results show relatively poor conformance in terms of realising the types of uses intended by land use regulations. The study area consisted of 24,78934 unique sites in 2016. The vast majority of these sites (22,458 – 91%) did not change use in the preceding 20 years. Of course, land use regulations do not propose changes on all sites in the study area. The most likely sites for changes are locations where a site’s land use has regulations that expect further or different types of development (i.e. under or exceeding areas). However, only a small proportion of these types of sites saw any change. A total of 4,901 sites were classified as under or exceeding for the entire twenty year study period, and only 728 of these sites (approximately 15%) changed use. Just what proportion of changed sites ought to be expected after a twenty year period is a subjective matter that is perhaps best addressed by an appeal to

34 This excludes the greenfield sites of Springfield and North Lakes. Development in these centres is regulated through a complex array of preliminary approvals and masterplans which are not directly comparable to the other centres.

208 conformance with overall compact activity centre principles. Does a 15% change represent a sufficiently significant alteration to the urban form to make more compact centres? The answer of course depends on the nature of the changes themselves but, looking across all centres, the results from Chapter 5 suggest that this rate of conversion is insufficient. Such relationships between land use regulations and centre conformance are considered further in Chapter 7. Next, the conformance of changed uses are considered.

Where the use did change, the change was classified using the previously described categories of conforming, under developed, and exceeding. In all locations, development had a high degree of conformance with land use regulations with an average rate of conformance across all centres of 85%. Figure 50 shows the rates of development conformance by location, with inner and outer centres showing greater rates of under development compared to the middle ring centres. This pattern is also apparent when looking at the individual centres (Figure 51), where almost half of the outer centres have greater than 10% under development, compared to only two of the inner/middle centres.

Figure 50 - Development conformance with land use regulations by location, 1996 to 2016

Percentage of changed sites 0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 100.0%

Inner N=152

Middle N=1372

Outer N=810 Centre Centre and name location

Conforming Under developed Exceeding

35 These figures exclude 87 sites which were under construction at the time of the 2016 GSV observation. As the final form of these sites were unknown, they could not be categorised by conformance.

209

Figure 51 - Development conformance with land use regulations by centre, 1996 to 2016

To better understand the nature of the development being undertaken, the land use database can be used to identify the type of use change for the different categories of conformance (Figure 52). The overwhelming majority of use change has been in the form of residential development. However, the type of residential use change varies by location, with inner residential changes primarily composed of higher density dwellings, while middle and outer areas showed a larger proportion of low density dwellings. Outer areas also differed by having a far greater proportion of changes to non-residential uses.

210

Figure 52 - Type of use change by location, 1996 to 2016

Percentage of changed sites 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Inner

Middle Location

Outer

Low density res Higher density res Commercial Bulky goods retail Industrial Under construction Others

The relatively high proportion of low density development sites may perhaps seem initially concerning in terms of conformance to compact city ideals. However, 85% of the low density development is occurring in locations that are regulated for this type of housing, making it highly conformant to land use regulations. Regional policy for activity centres does not dictate that all low density housing areas are to be replaced, and local governments are free to select the areas they deem most suitable for higher density housing. As shown in section 6.2, all centres are positively performing in this respect. An increase in conformant low density uses such as this is therefore not necessarily of concern in terms of an evaluation of conformance to regulations.

The remaining 14% of low density development that was categorised as under developed, made up more than 73% of all of the under-developed sites in the study area. This is potentially problematic for the implementation of compact activity centre policy, as newly constructed houses on sites identified for more intensive uses effectively blocks further development for the lifespan of the house; a period typically of at least two decades. This form of development also usually involved the subdivision of a larger lot, further fragmenting land ownership and thereby making future re-development more complicated. It should be noted that planning schemes typically regulate a maximum development intensity, and rarely require a minimum intensity. Planning schemes are drafted so that detached dwellings are almost always permitted when desired in low-medium and medium density zoned areas. Although it is possible to change planning schemes to more strictly control lower density uses in these situations (such as where some planning schemes are drafted to restrict new detached housing in high density zones), it raises questions of exactly where the line for a minimum intensity of development should be drawn, and whether a city wishes to potentially forsake some

211 development now in preference of a more intensive development that may never eventuate. It is also the local government that must face the inevitable call to justify the situation where a property owner is told they cannot build another house in an area that has been characterised by detached housing for decades, but that an apartment building or block of flats would be acceptable. Justifying such a position by appeals to sustainability related policy may have little charm when there is low public awareness of planning policy and mixed opinions of higher density development (Ruming, 2014). Additionally, such a decision effectively prevents “mum & dad developers” realising some development potential from their land as apartment construction requires specialised builders and more complex financing that is typically only accessible by more established developers. Imposing regulations that limit under development may therefore not only be unproductive, but also prove to be politically unpopular. Considering the overall level of under-development is relatively small, it is questionable whether action to limit under-development would overcome these extant issues and help implement compact city policy.

Sites where development exceeded what was envisioned by land use regulations were very rare, representing approximately 4% of changed sites. This type of non-conformance may be reflective of Queensland’s “performance based” planning system, which enables development to be approved contrary to land use regulations based on a qualitative assessment of its merits. Such an assessment includes aspects related to whether the use conforms to broader strategic intent such as regional level policies. Understanding the nature of the use of these types of non-conformities is therefore necessary to determine if they are likely to be problematic to the implementation of compact activity centre policy.

Only 3 of the exceeding sites consisted of uses that would not be expected by the regional activity centre policy; a warehouse in a non-industrial zone, and two contractors’ yards, all located in outer centres. In inner and middle locations, exceeding sites were almost entirely residential (100% for inner and 83% for middle). Residential non-conformances naturally result in greater amounts of dwellings than would otherwise have been provided and are therefore not considered contrary to overall policy aims as increasing overall population density is a core objective of activity centre policy. The exceeding sites in outer areas had a more diverse array of use types (Figure 53). The commercial category is primarily based on the development of strip mall and “big box” style of developments. Although some issue may be found with the degree of car dependency associated with this type of use, commercial uses of this nature are overall consistent with activity centre policy to provide employment and services and are therefore not considered to contrary to the regional policy. Likewise, bulky goods retail sites also have a place in activity centres according to regional policy, as do the

212 service industry (small repair shops, service stations etc.) and mixed-use industry (mixture of office/retail and industry) uses that predominantly make up the industrial category.

Figure 53 – Percentage of exceeding sites, by type of use, in outer locations, 1996 to 2016

12.2%

26.8%

19.5%

17.1% 24.4%

Higher density res Commercial Bulky goods retail Industrial Others

6.4. Discussion and conclusion

The key implementation mechanism for compact activity centre policy involves influencing local government land use regulations to be supportive of the types of uses intended by the policy. For the first 10 years of compact activity centre policy local government involvement was voluntary. From the gazettal of the first South East Queensland Regional Plan in 2005, local governments were required to consider the regional plan when making and amending their planning regulations. However, even under this more coercive system, the spatial extents of the activity centres are imprecise and the exact nature of the level of consideration that must be given to activity centre policy is not specified beyond some expectations of residential densities. Local governments therefore have maintained significant levels of discretion in how they may interpret and apply activity centre policy to their nominated centres, with this discretion being subject to final approval by the state government minister for planning. For the activity centre policy to be performing (as previously defined), there should be evidence of the policy being used when making decisions on matters relevant to centre policy. Due to the 20 year study period timeframe, and the number of local governments within the greater Brisbane area, it was not possible to analyse the decisions themselves. Instead, changes to land use regulatory documents that align with compact centre policy are considered as an indicator that the policy was used to inform local government decisions, and to therefore determine whether the policy was performing. This is considered a viable approach as changes to land

213 use regulations can be empirically investigated over time, and also represent a local government decision of some significance; a decision to change land use regulations has long term implications for a city’s future land use pattern, gives rights in land, is legally binding, and requires opportunities for public involvement.

The contents of the regulatory documents were investigated to identify direct references to the regional plan/s and the policy for compact activity centres. The inclusion of the regional policy terminology signifies that the regional plans were used in the formation of the regulations, and that the policy is therefore performing. The majority of centres in the study area included such references in the 1996 series of regulations, and by 2016 all centres were directly referenced as regional activity centres. These references proved to be more than lip- service to regional policy, with the regulatory controls themselves also showing evidence of decisions supportive of activity centre policy. The land use regulations were assigned development intensity scores for various types of land uses and applied to all properties in the study area. Changes to the intensity of permitted land uses were then considered over time and compared with the types of changes that would be expected of land use regulations that are aligned with centre policy. This analysis showed that the regulatory controls changed as expected, especially in terms of being permissive of more intensive residential and commercial use types. These results demonstrate that local governments are taking up the substantive aspects of activity centre policy and incorporating them into their land use regulations as intended. The final aspect of local government decision making in relation to urban development occurs through the application of the land use regulations, where a decision is made to approve or refuse a change or intensification of a use on individual development sites. Where uses did change, they proved to be highly conformant to land use regulations and regional activity centre policy. This shows that local governments are using the regulations as intended, to make development decisions to permit uses that align with the regulations and activity centre policy.

For the aspects of regional planning policy that local governments have direct control over (i.e. making and amending land use regulations and the application and enforcement of these regulations), local government decisions show strong evidence of performance. It is the aspects that local government typically do not directly control, namely the actual development of land, where policy intent and reality diverge. The results from Chapter 5 show that the nature of land use changes in the centres are aligned with compact activity centre policy, but in many centres the quantum of use change is insufficient to make material differences to compactness. This is further confirmed by the results of this chapter which show low rates of conversion of sites that planning regulations intended to be developed for different or more

214 intensive uses. The overall situation is one of strong performance of regional policy implementation but mixed to poor conformance with the intended outcomes. Local governments are acting as intended to accommodate compact activity centre policy into their own strategic planning and land use regulations and are applying these regulations as intended. This action alone however is clearly insufficient to deliver compact centres, which highlights that changing land use regulations (even though it is a precondition to enabling development), is inadequate as an implementation mechanism.

What factors account for the variability in conformance between activity centres? The next chapter seeks to address this question by examining the relationship between centre intensification and a range of potential explanatory factors.

215

7. Compact activity centre progress and its relationship to possible explanatory factors

Although popular in both public and planning imagination, the practicality of implementing Australian policies for compact activity centres has been questioned since their inception (section 3.2). Many of these critiques were predictive in nature, however some subsequent studies have tended to confirm these initial concerns with implementation (Chhetri, et al., 2013; Day, et al., 2015; Newton & Glackin, 2014; Phan, et al., 2009). To date, the majority of these studies have focussed on policies for metropolitan Melbourne. Chapter 5 expands this body of research to the case of greater Brisbane and finds similar issues with the implementation of regional activity centre policy. After 20 years, few of the designated activity centres conform to their planned outcomes. Chapter 6 provides a comprehensive analysis of land use regulations to determine the extent to which the policy performed in influencing local governments’ decisions to change land use regulations and approve development. The research found that the policy is performing well in these terms and that local governments are universally making regulatory decisions supportive of the regional policy. Yet, this has clearly proved insufficient to materially affect the planned outcomes. Section 3.2.2 reviewed explanations for this lack of implementation in other locations and identified factors relating to property and socio-economics, transport, planning policy, and existing urban forms. This chapter investigates the strength of associations between these factors, and the degree of centre intensification for all activity centres in greater Brisbane. The chapter firstly uses correlation coefficients to determine the strength of the relationships between centre intensification and the possible explanatory factors. This first part of the analysis finds that all groups of variables had moderate relationships to compact centre outcomes, except for measures of changes to planning policy. The second part of the chapter uses partial correlations to control for the potential influence of other factors. The results from this analysis show that property and socio-economic factors maintained the strongest relationships with centre intensification when controlling for the other variables. The existing intensity of employment in the centres also proved to have an independent relationship to centre intensification. Factors of distance and public transport were often better explained by other variables, whilst the extent of changes to planning regulations showed no relationship to centre intensification. These results are consistent with other research as well as with the findings from Chapters 5 & 6. The results indicate that the regional policy for activity centres lacks an adequate implementation mechanism and that this has broader implications for the role of planning in achieving more sustainable urban forms.

216

7.1. Relationships between compact centre intensification and factors

Section 4.2.4 outlined a method using statistical correlations to determine the relationship between a measure of centre intensification (developed in Chapter 5), and a range of factors that are commonly discussed in the literature as being influential on the implementation of compact city policy. In this section, the strength of these relationships is described for factors relating to initial centre compactness, property economics, transport, planning policy, and socio-economics.

7.1.1. Initial compactness These variables measure the compactness (as defined by regional policy) of each centre in 1996. Four sub-measures for density, dwelling mix, employment, and mixed land use were developed from a range built form and demographic data as described in detail in section 4.2.2, and the resultant scores are described in Chapter 5 and Table 70, (p295). The overall compactness variable is the average score of these four sub-measures. By correlating these scores with the score developed to measure intensification in terms of compact city principles, it is possible to see if there is a relationship between the initial nature of a centre and whether it became more compact as displayed in Table 42.

Table 42 - Relationship between centre intensification and existing compactness in 1996

Variable Pearson's r Potential Variance (%r2) 1996 Compact Score .538* 28.9% 1996 Density Score .459* 21.1%

1996 Dwelling Mix .386 14.9% 1996 Employment Score .598** 35.8%

1996 Mixed use score .269 7.2% *. Correlation is significant at the 0.05 level (2-tailed). **. Correlation is significant at the 0.01 level (2-tailed).

The results reveal that centres that were more compact in 1996 showed greater increases in measures of centre intensification as defined by regional policy intention. The results for the sub-scores reveal that of the overall compactness score the existing employment and density scores had the strongest relationship with intensification. These relationships are consistent with the predictions of Birrell, et al. (2005) who claim that the existing nature of the centres, particularly in terms of their employment concentration, will be a key influence in the realisation of activity centre policy.

217

7.1.2. Property factors The property variables are based on initial unit prices in 1996 and their relative and absolute change through to 2016. Unit prices were selected over housing prices as unit development is more aligned with the type of residential growth envisioned for activity centres. Section 4.2.4 discussed how construction costs are the largest component of apartment development and vary little by location. Development would therefore be mostly likely to occur in locations where unit prices are higher. Table 43 shows the strength of correlation between these variables and centre intensification.

Table 43 - Relationship between centre intensification and unit prices

Variable Spearman's r Potential Variance (%r2) Unit Price 1996 .663** 44.0% Absolute Unit Price Change .552* 30.5%

Perct Unit Price Change -0.238 5.7% *. Correlation is significant at the 0.05 level (2-tailed). **. Correlation is significant at the 0.01 level (2-tailed).

Centres where unit prices were higher in 1996, and centres which had larger absolute increases in unit prices over time, are positively related to centre intensification particularly in terms of existing prices. The weak relationship with relative unit price change is consistent with the reasoning for variable selection, as the required profit to undertake development is ultimately an absolute figure. Most centres experienced price increases over the twenty year period, and where unit prices were initially low, the relative increase was sometimes large. The unit price change variable must be interpreted with some care as it is not independent of the measure for centre intensification. Is the relationship due to increasing unit prices making development feasible? Or is the inverse true, in that feasible new developments result in higher unit prices? These questions cannot be answered by correlations of this nature. The existing unit price in 1996 however, is independent of the measure of intensification. This relationship is consistent with the numerous authors who look to explanations of property economics to explain compact city policy implementation (Birrell, et al., 2005; Brewer & Grant, 2015; Bryant, 2013; O'Connor & Healy, 2004; Rowley & Phibbs, 2012; Searle, 2004, 2010; Troy, 1996).

7.1.3. Transport factors The transport variables include the SNAMUTS composite index for public transport accessibility (SNAMUTS, 2016), to measures the accessibility and quality of public transport services in each centre. The road distance from each centre to the CBD is included to give some measure of motor vehicle travel, as well as a general proximity variable. As discussed in section 4.2.4, there is some debate regarding the influence of public transport on

218

development, with existing research from Melbourne showing little relationship between public transport services and the location of more intensive residential development (Newton & Glackin, 2014; Phan, et al., 2009). Searle (2010) believes that distance-based issues are affecting SEQ activity centre policy implementation, where too many centres are nominated in outer areas with insufficient “latent demand” for the types of development envisioned. The results in Table 44 are supportive of Searle’s position, with intensification decreasing as the road distance from the CBD increases. However, they also show a positive relationship between public transport accessibility and intensification. This raises an issue of whether good public transport is causing this relationship, or if it is more a case of good public transport being located in inner areas, and it is this distance from the CBD that is influencing the relationship with intensification more strongly than transit. The results from individual correlations cannot determine which is which36.

Table 44 -Relationship between centre intensification and transport factors

Variable Spearman's r Potential Variance (%r2) SNAMUTS Composite .469* 22.0% Road Distance to CBD -.513* 26.3% *. Correlation is significant at the 0.05 level (2-tailed).

7.1.4. Planning policy factors This section reports the results of correlations between measures of planning policy and centre intensification. The variables in Table 45 use the previously developed Development Intensity Scores (DIS) (Table 10, p110) from 1996 and 2016, to measure the change in development intensity permitted by land use regulations for different use types. The relationship between these variables and centre intensification are all weak. Changes in bulky goods retailing regulations show the strongest relationship, however this is most likely explained by the large increase in bulky goods land in Chermside and Upper Mount Gravatt (Figure 43, p197). If these centres are excluded from the correlation, the relationship becomes much weaker (r = 0.102).

36 See section 7.2 where the relationships are re-examined whilst controlling for other variables.

219

Table 45 - Relationship between centre intensification and planning scheme change factors

Variable Pearson's r Potential Variance (%r2) Change in residential zoning -.062 0.4% intensity - 1996 to 2016 Change in commercial zoning -.044 0.2% intensity - 1996 to 2016 Change in industrial zoning -.214 4.6% intensity - 1996 to 2016 Change in bulky good retail .296 8.8% zoning intensity - 1996 to 2016 No statistically significant relationships

As these variables measure a change during the same period as the change recorded by the dependent variable, it could be argued that it cannot be determined if land use change is altering the regulations or if the regulations are impacting the land uses. The results from section 6.3 however show that the majority of land use change is conformant with the land use regulations in effect at the time of change. This confirms that zoning changes are preceding land use change in most instances, and that the measures of planning scheme change are therefore mostly independent from the dependent variable.

Table 46 shows the results from the other planning policy variables. The first variable measures the proportion of sites in each centre that had a land use that was “under developed” (section 6.3) according to the applicable land use regulations in 1996. These sites are considered as areas that have yet to reach their development potential and were included here to determine if a relationship exists between the amount of such sites, and centre intensification. The results show a relatively weak positive association. The second variable shows how local government strategic policy reflected regional planning policy for compact activity centres. This is measured on a non-continuous ordinal scale as described in section 4.2.4, and highlights the early adopters of compact centre policy. There is a statistically significant positive relationship with this scale and centre intensification. This result should be interpreted with caution however, as these categories are almost entirely aligned with individual local governments and could therefore instead be considered as a correlation between local governments and overall intensification.

Table 46 - Relationship between centre intensification and other planning policy variables

Variable Spearman's r Potential Variance (%r2) Proportion of Under .304 9.2% developed Sites in 1996

220

1996 Strategic Reference .562* 31.6%

*. Correlation is significant at the 0.05 level (2-tailed).

7.1.5. Socio-economic factors The final variable correlates the Index of Education and Occupation (ABS, 1996) with centre intensification. Table 47 shows a strong positive relationship so that areas with higher initial levels of educational attainment and occupational status are correlated with centres intensification.

Table 47 - Relationship between centre intensification and the 1996 Index of Education and Occupation (ABS)

Variable Spearman's r Potential Variance (%r2) 1996 IEO Band .650** 42.3% **. Correlation is significant at the 0.01 level (2-tailed).

This variable is likely to be closely correlated with other variables such as unit price and distance to the CBD, and the relationship is therefore reconsidered once these aspects are controlled for.

7.2. Relationships with compact centre intensification while controlling for the influence of other variables

The previous section correlated a range of commonly cited explanatory variables with a measure of change of centre compactness, for all nominated activity centres in the greater Brisbane area. The results showed a number of statistically significant correlations for factors relating to initial centre compactness, unit prices, transport, and socio-economics. Factors relating to changes in land use regulations showed weak relationships. In this section, the relationships between these factors are considered when controlling for the influence of other variables, and to determine which of the factors continue to show strong relationships centre intensification when these other factors are taken into account. Relationships between variables were examined using a correlation matrix. Variables that demonstrated moderate relationships with other variables were then correlated with the dependent variable while controlling for these cross-correlations.

Of the factors so far considered, three have been discarded from this subsequent analysis. The change in unit price variables were discarded as the direction of the relationship could not be determined (e.g. unit prices may be affecting intensification or vice versa). The strategic

221

reference variable was discarded as it could just as easily be indicative of the influence of local governments as it is of any influence associated with the timing to adopt strategic policy. The strategic components of the planning schemes also lack any effectuating device other than how it is used to develop and enforce land use regulations. The changes to these regulations are also captured in the other variables that directly considered zoning changes. Additionally, several of the 1996 local governments are now defunct, and some centres were administered by three different local governments over the study period. Local government areas will likely be cross-correlated with other factors such as distance, property prices, and socio-economics, and the small number of centres is too limited to reliably assess the possible relationships between nine local governments on a multinominal scale, and these other factors. Spearman rank order correlations were undertaken between all the remaining variables to determine the strength of the relationships present between them (the full correlation matrix can be viewed in Table 82, p324).

7.2.1. Initial compactness vs. other variables The correlations between the initial compactness scores and the other variables are shown in Table 48. As would be expected, the overall compactness score correlates strongly with its sub-scores. There are also some strong relationships between the sub-scores, particularly between the density and dwelling mix scores. The dwelling mix score is derived from proportions of different dwelling types and favours the centres that include mixes of high, medium, and medium-low density dwellings. It is therefore not surprising that this variable is strongly related to measures of residential density. The employment sub-score has the greatest independence from the other compactness measures. Controlling for the density, dwelling mix, and mixed use scores, there is a statistically significant relationship between the employment score and centre intensification (r = 0.49937). This is not the case for the other sub-scores. When controlled using the other sub-scores, the density, dwelling mix, and mixed use scores show only weak (< 10% PV), non-significant relationships with centre intensification.

Comparing the other variables with the existing compactness scores shows a pattern of weak to moderate relationships with the distance to the CBD: as the distance increases, the existing compactness (across all measures) reduces. The compactness measures were correlated with overall intensification controlling for distance. Once again, only the employment measure maintained a moderate relationship with intensification (r = 0.467), and the other measures

37 Significant at the 0.05 level (2-tailed) – note: as discussed in section 4.2.4, statistical significance is only relevant for interpretation to larger population. It is reported here as is convention, however the coefficients used are descriptive of the observations for all principal and major activity centres in the greater Brisbane area and therefore even non-statistically significant correlations are relevant when describing this population.

222 including the overall compactness measure, demonstrated only weak relationships with centre intensification.

The relationships between the compactness indicators and changes in land use regulations are one-way relationships; the change in zoning cannot influence the initial compactness, but the initial compactness could potentially have an influence on decisions to change zoning. The residential category in particular showed strong relationships with the density based compactness measures. Centres that were more dense, and had a greater mix of dwelling types, saw more intensive future residential and commercial zoning changes, and less intensive industrial zoning changes. Controlling for the degree of existing compactness saw relationships between overall intensification and changes in residential and commercial zoning intensity become stronger but in a negative sense, so that where zoning changes were more intense, the overall centre intensification reduced. It is possible, but unlikely, that less restrictive land use regulations caused reduced intensification. Instead, this relationship highlights how regulatory change alone is not sufficient to drive centre change.

Table 48 - Relationship between existing compactness indicators and other variables

1996 1996 1996 1996 1996 Mixed Compact Density Dwelling Mix Employment use score Variable Score Score Score

1996 Compact Score 1.000 1996 Density Score .863** 1.000 1996 Dwelling Mix .784** .802** 1.000 1996 Employment Score .475* .312 .177 1.000 1996 Mixed use score .767** .496* .496* .211 1.000 Unit Price 1996 .168 .240 .191 .426 .142 1996 IEO Band .286 .274 .109 .498* .263 SNAMUTS Composite .329 .342 .098 .337 .355 Road Distance to CBD -.373 -.566* -.322 -.324 -.243 Change in residential zoning .645** .794** .662** -.030 .278 intensity - 1996 to 2016 Change in commercial zoning .474* .414 .415 .080 .330 intensity - 1996 to 2016 Change in industrial zoning -.326 -.531* -.288 -.319 -.044 intensity - 1996 to 2016 Change in bulky good retail .155 .187 .124 .124 .035 zoning intensity - 1996 to 2016 Proportion of under developed .625** .360 .575** .254 .770** Sites in 1996 *. Correlation is significant at the 0.05 level (2-tailed). **. Correlation is significant at the 0.01 level (2-tailed).

The relationship between initial compactness and the proportion of under developed sites is strongest in the categories related to mixed use and dwelling mix. It is reasoned that centres that are more mixed will naturally have a greater diversity of different site types compared to centres dominated by only a few, large format business based uses. Because there are a larger

223

number of sites in more intensively used parts of the centre (i.e. the areas where future development is more likely to be proposed), these centres therefore will have a higher proportion of under developed sites compared to less mixed centres. When controlling for the initial compactness measures for mixed use and dwelling mix, the proportion of under developed sites shows no correlation to centre intensification (r = -.014).

7.2.2. Property, socio-economics and transport vs. other variables There were also moderate to strong relationships between the key variables of unit price, the SEIFA scores, public transport accessibility, and distance to the CBD (Table 49). These relationships are all positive except for road distance, which indicates lower values for the other variables the greater the distance from the CBD.

Table 49 - Relationship between property, socio-economic, and transport variables

Unit Price 1996 IEO SNAMUTS Road 1996 Band Composite Distance to Variable CBD

Unit Price 1996 1.000 1996 IEO Band .715** 1.000 SNAMUTS Composite .425 .814** 1.000 Road Distance to CBD -.530* -.743** -.736** 1.000

*. Correlation is significant at the 0.05 level (2-tailed). **. Correlation is significant at the 0.01 level (2-tailed).

Controlling firstly for distance, both unit price and the IEO continue to show moderate positive relationships with centre intensification, whereas the public transport measure displays a significantly weaker relationship. Likewise, when controlling for public transport accessibility both unit price and IEO have statistically significant relationships with centre intensification. These results are detailed in Table 50 and demonstrate that the factors of unit price and socioeconomics maintain relationships to centre intensification independently of distance and transport factors. The relationships between transport factors and intensification however are well accounted for by the IEO as these relationships become particularly weak when controlled for this variable. Also of interest is the relationship between public transport accessibility and intensification controlling for distance. The strong relationship between transit and distance supports an obvious conclusion that locations closer to the city have better transport accessibility. However, if distance is controlled for, public transport has only a weak influence on centre intensification.

224

Table 50 - Relationships between centre intensification and property, socio-economic and transport variables while controlling for each of these variables

Controlling for… Variable Spearman's r Potential Variance (%r2) Road Distance to CBD Unit Price 1996 .537* 28.8% 1996 IEO Band .467 21.8% SNAMUTS Composite .157 2.5% SNAMUTS Composite Unit Price 1996 .580* 33.6% 1996 IEO Band .522* 27.2% Road Distance to CBD -.281 7.9% Unit Price 1996 1996 IEO Band .336 11.2% SNAMUTS Composite .276 7.6% Road Distance to CBD -.255 6.5% 1996 IEO Band Unit Price 1996 .374 14.0% SNAMUTS Composite -.136 1.8% Road Distance to CBD -.06 0.4%

*. Correlation is significant at the 0.05 level (2-tailed).

When controlling for unit price and IEO, weaker relationships are revealed with centre intensification (Unit price r = .374 and IEO r = .336). There is an obvious link here; it would be expected that those with higher status employment and better educational attainment would have higher paying work and be able to afford to live in areas with higher property prices and conversely, that areas with higher property prices would be more financially accessible to those with better jobs.

The final cross correlation of interest is the initial employment score and its relationships with the property, socio-economic, and transport factors. Unlike the other existing factors (density, dwelling mix, and mixed-use), the employment score has moderate relationships with all of these aspects (Table 48). As previously discussed, (and unlike the other initial factors) initial employment continues to show a statistically significant relationship when distance to the CBD is controlled (r = .467). Controlling for public transport accessibility yields similar results (r = .466). Property and socio-economic factors however explain more of the initial employment score’s relationship with centre intensification. Controlling for these aspects results in a more moderate correlation with intensification (r = .388 controlling for unit price, and r = .337 controlling for IEO). Controlling for initial employment however, has less impact on the relationship between unit price/IEO and centre intensification (r = .568 and r = .520 respectively).

225

7.3. Discussion and conclusion

A number of other studies (almost entirely from Melbourne) have shown or predicted that regional activity centre policy is not working as intended, and a range of factors have been proposed to explain this phenomenon. This chapter describes the results of a correlation analysis between the amount of centre intensification determined in Chapter 5 and possible explanatory factors including measures of property prices, transport and distance to the CBD, socio-economics, planning policy, and the initial nature of the centre in terms of density, dwelling mix, employment, and mixed-use. Analysis showed strong correlations between initial property prices, socio-economic indicators, initial employment structures, and transport factors. However, there was also a significant degree of cross-correlation between several of these variables.

Once key relationships between the different variables are accounted for, the economic based measures of unit price and the SEIFA Index of Education and Occupation, proved to be the most resilient. These variables maintained moderate relationships with centre intensification even when controlling for distance, public transport, and various measures of initial centre compactness. Although there was a degree of inter-relationship between these two variables the overall relationship was one where, for the variables considered, centre intensification increased more in centres where the 1996 median unit price and 1996 IEO, were higher. Such a relationship is consistent with the property based explanations for compact city development; locations that have a sufficiently high unit price to make development viable, and a highly qualified population with good jobs able to afford them, satisfy the basic feasibility requirements necessary for private led residential development to occur.

Aside from the employment measure, the relationships with initial compactness indictors (Table 42) and centre intensification, proved to be more strongly related to their distance from the CBD; i.e. centres that are closer to the CBD tend to be denser, however it is the closeness that explains centre intensification more than the centre’s density score. The employment score however was different. This variable continued to show moderate relationships with centre intensification when controlling for both distance and transport factors. It was also the only initial compactness variable that showed moderate relationships to centre intensification when controlling for the other initial compactness factors such as density, and mixed-use scores. As described in section 4.2.4, the employment score is derived from measures of employment density, intensity, and plot ratio. The relationship between these factors and centre intensification is supportive of the prediction of Birrell, et al. (2005), that for the activity centre policies to succeed, the individual centre would need sufficient employment concentrations to draw demand for housing and commerce from the surrounding area.

226

The public transport accessibility measure showed some correlation with centre intensification, however when controlling for other factors this relationship became weak. This was particularly the case when controlling for distance, suggesting better public transport is not linked to centre intensification (a result similar to the finding of Newton and Glackin (2014) regarding infill residential development). However, this should be interpreted with caution. There is little differentiation in public transport accessibility in outer areas, with these centres all being ranked in one of the bottom two categories on the SNAMUTS index. Greater Brisbane does not have an outer centre with the same quality of public transport service as inner centres such as Toowong, so the data cannot determine whether improved public transport to this standard may translate into more intensive centre outcomes. Distance to the CBD relates to a number of the variables investigated, and the results for Chapter 5 consistently displayed key differences between inner/middle and outer areas, across most of the measures examined. Distance better explained centre intensification than factors such as existing density and dwelling mix, however controlling for unit price and IEO resulted in distance having a weak relationship with centre intensification.

The planning policy indicators at no point showed relationships of strength with centre intensification. The degree of conformance of land uses to land use regulations was shown to primarily be the product of the association between this variable and the existing dwelling mix and mixed use factors. Changing planning policy to enable more or less residential, commercial, industrial, or bulky goods retail areas also did not show a relationship to centre intensification. In fact, these aspects showed negative relationships when controlling for the initial compactness of the centres. Chapter 6 demonstrated that local governments are rarely approving types of development that fundamentally contradict their regulatory intent. Land use regulations can therefore be considered effective at preventing forms of unwanted development. In this way, the regulations are a precondition to development occurring. However Chapter 6 also demonstrated that over time, all local governments in the study area changed their planning regulations to be supportive of compact centre development, particularly in terms of residential and commercial uses. The results from this chapter show that the act of changing zoning alone had little to no relationship with actual centre intensification, and the achievement of the activity centre policy. This result shows that the primary implementation mechanism of regional planning policy for compact centres is ineffectual in the absence of a market for the types of development proposed by the plan.

As described in section 4.2.4, it is important to note that the analysis of only 19 centres, although representative of all principal and major activity centres in the study area, is naturally limited, particularly in terms of determining causation. The use of simple correlations as

227 described above can only point towards to the presence of relationships. Although some key relationships were identified even when controlling for other variables, the presence of other confounding factors may better explain the phenomenon. Caution should also be used in applying the observed relationships to a wider population; although many of the relationships are statistically significant, with only 19 samples their inferential power is low. None the less, the results presented here are consistent with common explanations for the lack of centre implementation in other Australian cities and adds further evidence for how commonly cited factors explaining compact city implementation relate to twenty years’ of development in the greater Brisbane area.

The next chapter concludes the thesis by drawing together key findings to discuss how this research relates to plan implementation evaluation and its methods. The example of greater Brisbane is also considered in terms of the role of activity centre policy in achieving more sustainable urban forms, the future prospects for this type of policy, and a path for future research to further understand centre policy implementation.

228

8. Conclusion and discussion of results

This section considers the research findings in terms of their implications for compact activity centre policy, theories and methods of plan evaluation, urban sustainability, and future research directions. The research found that few centres showed strong conformance to compact activity centre policy, even though the policy was being applied as intended by local governments. Instead, centre intensification was better explained by relationships to existing socio-economic factors and types of employment based land uses. These findings are discussed in the context of centre selection and policy implementation. The discussion introduces and contrasts two approaches used by different local governments; one approach focussed on placemaking as a city marketing strategy to attract investment, and the other involved direct government investment in commercial development. The discussion shows the latter approach has more promise and is better aligned with normative planning goals, but ultimately centre policy needs to better consider the practicality of selected implementation mechanisms and undertake more detailed assessments of potential centre locations before nominating them as activity centres. The research’s findings indicate a degree of resilience to change to the fundamental characteristics of centre form. This, as well a preliminary analysis of transport data, introduces doubts as to whether attempts to restructure the urban form in this manner can offer meaningful improvements to sustainability.

The discussion highlights how this research is a necessary first step in undertaking further conformance based evaluation that focusses on the influences of compact activity centre policy on sustainability outcomes; the overall objective of the policy. The discussion also addresses the role of conformance and performance based approaches to plan evaluation and finds that the two concepts are fundamentally related, and both offer important insights. The limitations of the deductive form of performance evaluation used in this research are discussed, showing that different forms of performance analysis would also be beneficial to further understandings of the topic. In particular, performance evaluation that examines the sociological and institutional factors that influence policy making and its implementation is shown to have potential in developing explanations for how and why policy makers make decisions in relation to plan implementation. As practitioners create plans and policies with the purpose of effecting change, evaluating the conformance of change to policy is also shown to be essential. Future research that expands to include activity centres from other cities, as well as “non-centre” locations, would enable the use of more sophisticated methods of analysis.

229

8.1. Overview of research results

Land use planning policies to encourage more compact urban forms are now ubiquitous in most OECD countries. With compact city policies present in every capital city, Australia is no exception. However, the nature of Australia’s sprawling conurbations led to the adoption of a particular form of compact city policy that goes beyond simply increasing residential densities. Instead, these policies also seek to disburse the primacy of the CBD through the development of a network of compact nodes across the conurbation. Most Australian cities had long had policies for commercial centres, however with increasing concerns relating to urban sustainability, these policies were adapted in line with compact city principles. This was done in order to create more sustainable urban forms. It is reasoned that having more people living near concentrations of diverse employment, services, and public transit opportunities would reduce automobile dependence and thus provide sustainability benefits. The development of activity centre policy in the greater Brisbane area followed this same trajectory. Regional planning policy for Brisbane has now included activity centres for more than two decades. This policy has been remarkably consistent during this period, most centres nominated in the initial regional planning policy documents continuing to be nominated as centres in the current iteration of the regional plan. The policy intent has also remained consistent throughout this period, with the policy calling for the development of activity centres to be characterised by higher residential densities, a greater diversity of dwelling types, and higher concentrations of mixed employment generating uses.

Despite the importance given to activity centre policies in regional planning documents, there have been relatively few studies that empirically evaluate either their implementation or effects. Most studies that have done so reveal that the implementation of activity centre policy has been disappointing (Bunker, 2014; Chhetri, et al., 2013; Day, et al., 2015; Newton & Glackin, 2014; Phan, et al., 2009). These studies often make use of data that is now 10 years old and which therefore miss recent changes such as the apartment construction boom that occurred in Australian capital cities over the past decade. Existing studies also only consider changes over periods of ten years or less and are almost entirely focussed on the case of Melbourne. The only empirical study of Brisbane’s activity centres that could be found provided a positive view its implementation (BITRE, 2013). However, this study used an approach that did not adequately account for the inherent issues with longitudinal census data, used inconsistent centre boundaries, and failed to differentiate between different types of nominated activity centres. The existing studies are also almost entirely focussed on plan conformance, i.e. evaluating planned outcomes against actual outcomes. This line inquiry is valuable, however planning evaluation theorists also call for studies of plan performance

230 which investigate how plans were used by decision makers. Combining an evaluation of performance and conformance gives plan evaluation further explanatory strength by enabling a determination of whether the observed outcomes were the result of the use of the plan, or due to other factors. Although a range of authors have outlined explanations for why compact city centre policies have struggled, few have looked specifically at activity centre policy or attempted to empirically link possible explanatory factors to actual policy outcomes. This research developed three research questions to address these gaps.

The first question sought to understand how greater Brisbane’s activity centres have changed in-line with compact city based metropolitan policy. By doing so the research has, for the first time, provided a thorough evaluation of activity centre conformance over a 20 year time period. It is also the first comprehensive evaluation of activity centre policy in the greater Brisbane area. Additionally, a new method was developed to overcome a range of data limitations that have restricted previous research, and which can potentially be applied in other locations to further expand the research scope, and the suitability of different methods of analysis. This method used Google Street View and aerial imagery to develop a detailed land use database. This database was then used as an auxiliary dataset to undertake dasymetric areal interpolation of census data, as well as to estimate employment and describe land use change. The data was used to make measurements of 15 different indicators across a range of factors aligned with the intended outcomes of activity centre policy. Measurements were taken at both the beginning (1996) and end (2016) of the study period and were also used to measure the degree of centre intensification during this period. Where feasible, centre intensification was also compared to baseline measures of intensification occurring in the broader urban area. The results show typically poor levels of conformance, with only two centres conforming to all aspects of the policy. An additional five centres conformed reasonably to key parts of the policy, with the remaining centres typically failing to show levels of intensification that exceeded the general level of intensification recorded in the broader conurbation. These results are consistent with the existing studies of activity centre conformance in Melbourne. Most non-conforming centres were located in outer areas and the non-conformance in these locations was primarily the result of a lack of development activity, rather than the development of the “wrong” type of uses.

The second research question sought to investigate how metropolitan planning policy for compact activity centres has influenced land use regulations, and how actual land use changes conform to these regulations. As regional level policy is rarely applied directly to individual development decisions, understanding if the policy was used as intended in local government regulatory land use plans is important to determine whether the poor conformance was due

231 to a lack of use of centre policy by local governments or the result of other factors. The approach used was deductive; if references to activity centre policy could be found in land use regulations, and the regulations themselves changed to reflect the intended policy outcomes, then it could be concluded that local government decision makers had used the policy to make land use planning decisions. To address these aspects, current and past local government land use regulations were examined in detail. A basic content analysis was firstly undertaken to determine if the regulations changed to textually reference activity centre policy. The results of this analysis demonstrated that over time, all local governments amended their policies to directly reference the roles of all activity centres. Next, the details of the regulations themselves were examined to determine if they had changed to enable the forms of development intended by activity centre policy. To do so, the regulations were coded into Development Intensity Scores (DIS) for different categories of land use. The DIS were then compared over time to determine if the regulations were enabling more intensive forms of development. The results showed that in the critical categories of residential and commercial use, the regulations for all centres changed in accordance with regional policy. Finally, the conformance of actual land use change was compared to the regulatory intent of local government plans. This revealed that the overwhelming majority of use changes conformed to the regulated intent, and only a handful of examples of non-conforming, non-centre development occurred. This result shows that local governments are using their land use plans as intended when assessing development and land use changes. Combining these various results presents a clear picture of strong plan performance; local governments are using activity centre policies when making or amending their planning regulations, and they are then using their regulations as intended in their development assessment processes. The activity centre policy’s poor conformance is therefore happening despite strong plan performance, indicating that the policy uses impractical implementation mechanisms.

The final research question sought to determine the relationships between commonly cited explanatory factors for poor activity centre conformance, and the achievement of compact activity centre objectives. The factors considered included property prices, transport and distance to the CBD, socio-economics, and the initial nature of the centres themselves. Also considered was the potential influence of changed land use regulations. Centre conformance was determined using the centre intensification score identified as part of the results of the first research question. This score was then statistically correlated with measures of the various factors to determine the strength of the observed relationships. Negligible relationships were observed between the degree of land use regulatory change and centre intensification. The strongest relationships were observed between socio-economic factors, initial unit prices, and

232 the initial employment intensity score of the centre. These results are consistent with property market based explanations of centre development.

8.2. Research implications

The research results identified a key disconnect between compact activity centre policy, and the selected mechanisms for its implementation. The designated centres were to become key regional nodes characterised by a range of higher density residential development and employment generating commercial, service, and administrative uses. This policy was to be implemented through updated regulatory land use plans that reflected these objectives, improved coordination of government services, and infrastructure delivery. Of these, state and local statutory land use planning systems were to be the “key delivery mechanism” of the regional plan (The State of Queensland, 2009, p. 7). The planning regulatory system is the one key aspect of the development cycle that the planners at both state and local governments have direct influence on. The state government, through its eventual formalisation of the role of regional plans, was able to require local governments to develop plans sympathetic to regional aims. The local governments maintained the authority to determine how and where these regional objectives would be aligned with the development rights associated with individual properties. Where there is market demand for the type of development envisioned by the regional plan, these regulatory changes shape the urban form by restricting or permitting the scale and type of use. But where no such demand exists, the regulations change without a corresponding change to the physical world. The research showed strong conformance between land use regulations and land use change, which demonstrates that aligning regulations with policy intent is clearly a necessary precondition for the development of the uses desired by the policy. However, the results from the three research questions reveal that the primary implementation mechanism of compact activity centre policy (making changes to the statutory land use planning system) is insufficient to drive change where a market for the types of development proposed by the plan is not present. And it was this issue that resulted in the poor conformance observed in the centres. Conformance issues were less the case of the “wrong” type of development occurring, but a case where the amount of development was insufficient to create the idealised compact centre form. The disconnect here is that the centre policy, in using the regulatory planning system to effect change, assumes that regulatory changes will somehow lead to the desired material results. The findings from the third research question however, demonstrated that there was no relationship between centre intensification, and regulatory change.

233

The research instead found relationships between a range of other interrelated factors, and centre intensification. The strongest and most resilient of these relationships were from unit prices, and the residents’ educational attainment and occupational status. The centres characterised by higher unit prices, and with more highly educated residents with skilled jobs, were more likely to have greater levels of compact intensification. Similarly, centres with existing arrangements of more intensive forms of employment based uses were also more strongly related to centre intensification. There was also some relationship with distance to the CBD, with centres closer to the CBD showing a greater degree of intensification. However, these centres also typically scored more highly in relation to the other factors previously described. These factors are related to property market explanations of centre development. If planning is expected to influence outcomes in a system dependent on privately instigated property development, were these factors were considered when the designated centres were initially selected?

8.2.1. Policy formation and selection of centres Section 2.2 reviewed historical policy documents to provide an overview of the development of the compact activity policy over time. Initial policy documents identified “candidate centres” (Figure 2, p35) primarily on the basis of the size of their retail catchments and transport accessibility. However, a key action of this early policy was for further work to be undertaken to identify suitable locations for major centres on the basis of a range of criteria including public and private transport accessibility, the location other centres, the availability of land for centre expansion, the existing size and function of the centre, and “…its capacity to increase the range and density of jobs and services” (The State of Queensland, 1993b, p. 48). A subsequent review of this initial policy work promoted a more cautious approach, with restrictions on out of centre development, and government infrastructure and employment provision, being “contingent” on a positive response to centre policy by private sector led development (The State of Queensland, 1994a, p. 2). This report went on to describe a preferred process for subsequent planning where the nominated centres are identified using criteria related to catchment size, “latent private and public sector support” and the “physical capacity for expansion and integration as a cohesive town centre” (The State of Queensland, 1994a, p. 13). These criteria have a strong focus on development feasibility, and specifically seek to determine whether a nominated centre could attract sufficient private sector development of both employment and higher density residential uses. However, with the release of the first officially endorsed regional plan in 1995, major centres continued to be identified only by population catchment and the number of existing jobs (The State of Queensland, 1995). A key action of this plan was to undertake a study to spatially identify the location of appropriate major centres. A subsequent scoping study to guide this work noted

234

the importance of market based mechanisms and that “the development of centres is a function of demand for office floor space, retail floor space and other uses which in turn are a function of population and socio-economic factors” (Planning Workshop Australia, 1996, p. 3). However, the major centre study never materialised. Major centres continued to be identified only by catchment and job numbers, and the key action to undertake a study to locate the centres remained in all iterations of the regional plans until the release of the first South East Queensland Regional Plan in 2005 (The State of Queensland, 1998, 2000, 2005c). With this plan, came the first officially adopted spatial recognition of all activity centre types. Unfortunately, only limited background information was released with the plan and no further justification for these locations could be accessed through public records. The 2005 centre locations (Figure 6, p42) bare remarkable resemblance to the initial candidate centres of the 1993 policy papers, as well as the centres listed in the 1996 scoping study, with only the centres of Wynnum and Redcliffe being previously unmentioned additions to the centre hierarchy. These centres from the 2005 plan remain as the nominated centres in the current iteration of the regional plan and show a continuation of the initial regional policy for centres since its inception.

Despite calls for the regional plan to better consider the market conditions necessary to implement the plan from both internal and external policy makers, and plan submitters (The State of Queensland, 2005a, p. 122), the endorsed plans continued to identify centres that lacked the basic criteria necessary for private development to occur. Based on the documentation available, the nominated activity centres appear to have been selected without consideration of the key factors linked to centre intensification; existing employment structure, unit prices, and socio-economic conditions. The key implementation strategy of altering statutory plans to be more supportive of compact centre based uses, was therefore an unsuitable mechanism by which to implement the policy in the selected centres. The patchy conformance to planned outcomes was primarily due to some centres happening to have conditions suitable for development to proceed, rather than being due to the considered application of regional planning policy. Absent these conditions, changing land use regulations to encourage new development represented more a case of wishful thinking than planning for actual change.

8.2.2. Other implementation mechanisms The regional plans also called for implementation to be driven by the development of key infrastructure, and direct government support through locating regional offices and other public sector employment opportunities in activity centres. Infrastructure objectives typically related to calls of further planning and investigation, although other proposals such as

235 improvements to bus links between centres, and the development of bus interchanges were described in more concrete terms (The State of Queensland, 1995). The release of the South East Queensland Infrastructure Plan and Program in conjunction with the 2005 SEQRP provided more direct linkages between the planning policy and infrastructure planning (The State of Queensland, 2005b). Although this plan acknowledged the importance of the activity centre concept, there was little direct link between centres and infrastructure provision. Major transport projects related to upgrades to the Ipswich motorway and other road projects not specifically related to centres policy, and busway proposals such as extensions to Capalaba and Springwood (which at the time of writing had yet to materialise). There were also proposals to upgrade the Ipswich and Prince Charles Hospitals. A number of these projects proceeded. Notable centre specific transport projects were the South East Busway connecting Upper Mount Gravatt, the development of bus interchanges in numerous centres, and a train line extension to Springfield and Redcliffe. The hospital upgrades also proceeded in Ipswich and Chermside. It is not clear how much these changes assisted with overall compact centre development. The changes had direct effect on relatively few centres and even after the transport upgrades, the vast bulk of greater Brisbane’s centres were categorised as having either poor, minimal, or less than minimal public transport services (SNAMUTS, 2016). The results from Chapter 7 failed to find a clear relationship between public transport accessibility and overall centre intensification when controlling for other variables.

The institutional public investment in hospital upgrades however did yield results, with increases in hospital related employment representing a large proportion of the employment growth in these centres. Likewise, institutional support in the form of government services being relocated or expanded in centres also made a measurable difference to office employment levels (Table 64, p291). Office development in centres such as Logan Central for example, were almost entirely from the public sector with the development of an expanded council administration centre, a federal government Centrelink office, and state government health facilities. Ipswich also saw large employment gains from state and local government office relocations and redevelopments, and other institutional upgrades such as their new cultural precinct and court house. However, this employment growth failed to influence the private market for other forms of development, particularly residential uses. Large increases in government led employment such as that observed in Ipswich, were the exception rather than the rule, and the overall principal changes in employment were the result of retail based development (Figure 16, p147). There is a finite amount of demand for new and relocated government services. With 21 principal and major activity centres in the greater Brisbane area (and 16 additional centres nominated in the broader region), and a lack of further guidance as to which of these centres ought to be prioritised, there is limited capacity to use this

236 mechanism to provide the concentrations of employment necessary to stimulate compact forms of development.

So, where does this leave the policy for compact activity centres, particularly in the nominated centres that lack the existing employment concentrations and socio-economic conditions necessary to stimulate the required forms of development? Strategic plans and land use regulations have changed to reflect their activity centre status, but where the conditions for private or government led development are absent, it is likely to be a long time before any of the purported benefits of more compact centres are realised. Local governments are pursuing different approaches in an attempt to implement their centre plans. Of these, the approaches of Logan and Ipswich City Councils are illustrative of current attempts.

Logan’s marketing, placemaking, and development attraction approach Logan City Council (LCC) has adopted an approach which links its strategic planning activities with marketing driven economic development. Central to this approach has been the hosting of a number of “summits” to help guide the implementation of their centre planning. The end result is the identification of potential “place making” projects intended to stimulate development. To date, summits have been held for a number of regional activity centres including Springwood, Beenleigh, and Logan Central, as well as the specialised centre at Meadowbrook and the major centre of Jimboomba38. Previous planning for Springwood had taken up the regional planning intent for the centre with some enthusiasm, with supportive land use regulatory changes in both the 2006 and 2015 Logan Planning Schemes, as well as in policy documents such as the Greater Springwood Master Plan which envisioned Springwood as a thriving commercial centre, packed with glass clad, high rise office and residential towers (Figure 54).

38 Jimboomba is located outside the 35km greater Brisbane radius and was therefore not included as a part of this research.

237

Figure 54 - Artist's impression of the future of Springwood (Logan City Council, 2009)

Despite these intentions, Springwood showed some of the most limited progress in achieving compact centre objectives (section 5.5.1) and provides a good example of the nature of LCC’s centre summits and their approach to stimulating development.

The Springwood Summit was held in October 2016 on the top floor of Springwood’s, mostly vacant, and only high rise commercial building, and promised to bring “national and international perspectives on unlocking the economic and placemaking potential for Springwood and the City of Logan.” (Section 10.8). In line with the event’s tagline of “place, investment, and collaboration”, LCC reported that “key government, industry, business owners and landowners collaborated to unlock the economic and placemaking potential for Springwood” at the summit (Logan City Council, 2016). Speakers for the event hailed from urban think-tanks, development corporations, and design and planning firms. The agenda was heavily focussed towards creativity, placemaking, and bringing government and private sector interests together to implement shared visions. Acknowledging that the planning for Springwood over the past two decades “hasn’t really worked”, Logan’s (then) Mayor saw the council’s role as one to “generate excitement and investment around Springwood” (Orr, 2016). This marketing focus is unashamedly prominent in the event promotional material which proclaims “developers are provided with all they could ask for… Springwood is ready, willing and waiting to embrace its business and commercial development potential and become South East Queensland's newest central business district” (section 10.8).

238

In the presentations, “collaboration” (particularly between government and private development interests) was touted as the key to centre success. Some participants raised fundamental questions around the implementation of the Springwood vision such as what infrastructure will be put in place to support a CBD scale development, who will pay for it, and why businesses and residents will want to move there. However, the concluding session deferred these issues in favour of ideas such as marketing based actions to exploit the centre’s supposedly high levels of cultural diversity, promoting its existing light industrial areas to make it a centre for high performance vehicles, and delivering improved street scaping. Other ideas focussed on further relaxations of council regulations where development controls are suspended in a “regulation holiday” to promote investment. The end result of the summit was a list of actions to help promote and develop the centre. These included projects to “advocate” for and investigate some major transport projects such as widening the M1 and delivering a busway extension, new pathways, street landscaping and street art, and upgraded public spaces (Logan City Council, 2017a). The most immediate outcomes however involved the creation of a “Springwood Economic Development Zone” with less stringent car parking and height controls, as well as discounts and deferrals on infrastructure charges for desirable forms of development (Logan City Council, 2017b).

Both the event and its outcomes show the LCC approach involves merging the roles of strategic land use-planning and city marketing. The Springwood summit had a clear purpose: to implement the council’s compact city visions for Springwood. The chosen mechanism was a marketing event to create interest among stakeholders associated with the development industry. Ideas generated from the event were framed in the same manner; undertake placemaking and relax regulations to attract investment.

Such approaches to city development see cities enter into competition with one another to attract private investment capital, and thus implement their plans. This is consistent with the views expressed earlier in the research (section 2.1.1) which critiqued the neoliberal approach of modern planning where cities under such a system assume the “…role of a courtesan hoping to seduce investors who will shower material blessing on her” (Sandercock, 2005, p. 326). This type of approach is contrary to normative goals of regional planning which seek to counter-act these competitive tendencies in favour of the orderly allocation of growth and resources to shape a cohesive urban form.

239

Ipswich Council becomes a property developer Ipswich City Council (ICC) took a more direct approach to generate development in the Ipswich city centre. As discussed in section 6.1, ICC was an early adopter of regional activity centre policy for its city centre and quickly moved to recognise its activity centre status in both its strategic plan as well as in its land use regulations. Even from these early stages, Ipswich has sought institutional investment, and used the initial policies to help justify and secure a new university campus39 (The State of Queensland, 2000). Subsequent planning for Ipswich also drew particular attention to the centre’s status as a regional activity centre, even incorporating this role directly into the vision of key planning documents whereby Ipswich was “to be the vibrant and prosperous regional activity centre for the western corridor of South East Queensland” (The State of Queensland & Ipswich City Council, 2007, p. 43). The focus on collaboration with institutional actors continued with this planning, which sought to deliver a series of “catalytic projects… to actively incubate, facilitate, stimulate and ‘kick start’ the revitalisation of the Ipswich Regional Centre” (The State of Queensland & Ipswich City Council, 2007, p. ii). Among these projects were actions to locate government offices in Ipswich, assemble land for development in public-private partnership, and to revitalise key sites in the Ipswich CBD. These actions took shape via the formation of an ICC owned development company, which borrowed $50 million dollars from the state government (via ICC) to purchase a large shopping centre site in the Ipswich CBD (Gardiner, 2009).

The new development company (Ipswich City Properties (ICP)) sought and selected a private sector partner to plan and develop the site, resulting in a scheme to construct a number of high-rise office and residential towers incorporating entertainment, major retail, and public spaces (Ipswich City Properties, 2009; Jackson, 2010). Stage 1 of the project advanced rapidly with the construction and sale of a high rise mixed-use office building (ground floor retail tenancies with offices above), made possible by the state government agreeing to lease 90% of the new building for 15 years (Cromwell, 2011). Lacking further institutional support, progress slowed and a subsequent 2013 development permit for stage 2 of the project was never commenced (Ipswich City Council, 2013; Google Street View, 2017). The site masterplan was later revised to exclude this tower and the residential components of the development, and instead provide a new council chambers with retail areas and open space (Ipswich City Properties, 2017). Amid allegations of corruption by the city’s mayor and senior executive, reports of the operation of ICP losing $80 million, and the state government sacking the city council and appointing a city administrator, a decision was made to dissolve ICP and

39 This is the University of Southern Queensland’s Ipswich Campus (formerly of the University of Queensland). The campus is approximately 2km from the Ipswich centre and was therefore not included within the study area extent for this research. Section 4.2 discusses the rationale for the selected centre extents in detail.

240

bring its operations back under the direct control of ICC (Johnson, 2018; Robertson, 2018). None the less, in recent statements ICC claims to be committed to completing the project (Howarth, 2019). During the study period, Ipswich also secured government investment for a range of other significant projects such as a new courthouse, a large hospital upgrade, an upgraded police station, and a new civic precinct with art gallery and office spaces. These, along with the government office relocations associated with the ICP project and the development a new big box shopping centre, resulted in Ipswich scoring 4th out of all activity centres in terms of employment intensification during the 20 year study period (section 5.3).

ICC’s direct approach to investment and partnership with the state government has resulted in tangible changes to the centre, as well as increases in skilled employment. This has not been without risk though, and the commercial realities of the Ipswich market showed that ICP’s development was entirely dependent on public sector support. It is unlikely that similar levels of support could be maintained across all of the designated activity centres. As an historic town of some regional importance, Ipswich already ranked highly in terms of its levels of existing employment and mixed-use, and these existing businesses and institutions helped to make the case for subsequent investment (Cromwell, 2011; Ipswich City Properties, 2009). The case for similar investment in centres lacking these starting attributes would be a harder prospect to sell to either public or private sector entities. Ipswich’s commercial development also failed to generate demand for more intensive residential development. This is a key consideration for compact centres as without the residential component, many of the purported sustainability benefits of compactness, are no longer achievable.

8.2.3. Planning in a “parallel universe” Based on these examples, neither the Logan nor Ipswich approaches to centre implementation bode particularly well for the future development of the current array of centres. Using direct government investment as an implementation mechanism would require either huge investments of capital, or a more select number of centres in which to concentrate this investment to gain the maximum impact. Even then, prudence would require a careful consideration of the benefits of such actions compared to the costs, especially as the case for improved sustainability of more compact centres remains to be demonstrated in an Australian context (see section 8.2.4 below). If private sector led development is to remain as the key method of centre implementation, then regional planners need to more closely consider the existing nature of the locations they designate as centres to ensure these areas offer the conditions necessary for the desired form of development to proceed. This requires a thorough analysis of existing land availability, the nature of existing residential and commercial land

241 uses, socio-economic factors, and market feasibility analysis. The selection of the current array of centres apparently lacked such considerations.

Even if such considerations had been made, the suitability of planning to meet the market is questionable. Land markets are notoriously difficult to predict and are subject to sudden and unexpected changes such as those brought on by the Global Financial Crisis of 2007 (Krugman, 2009). Consumer trends for housing purchases can change rapidly (Wu & Brynjolfsson, 2015), sometimes on timeframes shorter than those required to complete the drafting and approval of comprehensive land use plans such as the SEQRP. Writing policy to suit the fickle ebbs and flows of the market entrenches neoliberalised forms of planning (Olesen, 2014; Taşan-Kok, 2012) in lieu of the more normative strategic approaches proposed by planning theorists (Albrechts & Balducci, 2013) and by the regional plans themselves which seek to achieve a wide range of value based outcomes that are unlikely to be delivered through market responses alone. Relying purely on private market based development is also an exclusive approach that potentially precludes implementation of desirable planning outcomes, such as the sustainability benefits supposedly inherent in compact activity centres, in the car dependent outer centres which lack the demand for private sector development. The solution of better incorporating market preferences into plan making is therefore problematic both normatively and practically. None the less, it is clear that planning policy needs have more realistic implementation mechanisms if there is an expectation of material achievement of the plan’s objectives.

The results of this research unfortunately contribute to a disappointing pattern of Australian strategic land use implementation failures, whereby grand visions were developed absent the presence of feasible mechanisms for their implementation. Australia has a long history of large scale policies that have attempted to manipulate urban growth both within and away from cities, typically with poor results (Davidson, 1997; Jain & Courvisanos, 2009; Lonsdale, 1972; Simons & Lonergan, 1973). Even forty years ago, scholars argued that attempts to centrally plan large scale population and employment relocations in a society that “…places an overriding premium on economic productivity, efficiency, and growth”, were working against economic forces that were “…too powerful and too fundamental to be overcome by the kind of efforts that governments have been willing to take” (Lonsdale, 1972, p. 328). Twenty years later, McLoughlin (1992) demonstrated that planners’ efforts to reshape Melbourne according to a predetermined pattern of commercial centres had failed to materialise, leading to the conclusion that planning had achieved few of its more strategic objectives and instead succeeded primarily in maintaining suburban forms of development through the enforcement of prescriptive development controls such as building setbacks. The rise of compact city policy

242 is another example of planners once again seeking to resolve urban issues through fundamentally reshaping broad-scale urban forms. In the incipient days of these policy movements Breheny (1997, p. 215) cautioned that implementing the compact city was likely to prove highly challenging due to the “daunting” task of “…reversing the economic geography that underpins demographic geography”. These challenges are even greater when considering the compact activity centre concept, which requires large scale urban development trends to be channelled into relatively precise geographic areas that are often far removed from locations with the greatest economic demand.

This was also a key conclusion of a prophetic paper by Birrell, et al. (2005, p. 48) in relation to Melbourne’s activity centre policy which they believed had failed to “…keep in perspective the difficulty of changing land-use patterns sufficiently to meet the broad objectives of [the] strategy, given the limited tools available to the current planning system.” Birrell, et al. (2005) predicted that the Melbourne activity centre policy was destined to fail for a number of reasons. They argued that the policy disbursed development across too many centres, and that the nominated centres were too poorly defined in both spatial and typological terms. These aspects were observed in the greater Brisbane centres which were too numerous to concentrate direct investment and which did not distinguish between differences in centre functions in their selection. Chapter 7 demonstrated the importance of these considerations, with the initial centre employment structure showing a strong relationship with future intensification. Birrell, et al. (2005) also predicted that employment growth, particularly higher order office based employment, was unlikely to materialise in activity centres as planned as these jobs were most likely to be concentrated in locations nearer the CBD. They reasoned that existing employment patterns suggested activity centres were more likely to attract a modest amount of lower order retail, accommodation, cafe and restaurant based employment. This prediction was borne out by the experience of greater Brisbane’s centres which demonstrated that the overwhelming majority of employment growth was in the form of such low order employment uses (Chapter 5). Doubts were also raised about the feasibility of higher density residential development in Melbourne’s centres with Birrell, et al. (2005) predicting that such development was unlikely to occur due to a lack of suitable development sites within centres, and that past patterns of residential development suggested that Australian workers were unlikely to demand homes in immediate proximity to their employment. This prediction proved to be accurate in greater Brisbane, however it was more the result of the latter phenomena over the former. In greater Brisbane, higher density residential development most typically replaced detached dwellings; a use type of which there was no shortage within the nominated extent of each centre. Although some local governments’ seeming reluctance to rezone detached housing for high density uses (e.g. Brisbane City Council – Chapter 6) could

243

be seen as one such limitation to site availability, on the whole large areas of land were available for development. The greater issue appears to be a lack of market demand for these housing types. The previously described case of Ipswich is illustrative of this, where higher density residential growth failed to materialise even in the presence of higher order employment growth.

Forster (2006) notes that activity centre policies appear to be conceived in a “parallel universe”, where planning intent exists in a realm separate from the realities of the Australian urban form and development processes. The results of this research do not go so far as to conclude that the planning system is incapable of achieving its intended outcomes. However, the results do raise doubts and demonstrate that the impact of regulatory change and piecemeal infrastructure/direct government investment were insufficient to stimulate the development necessary to achieve more compact activity centres. The ultimate purpose of the centres policy is justified by the reasoning of achieving ecological sustainability. Although the activity centre policy was not expected to achieve this result alone, it was intended to have positive impacts on this aim. This research did not seek to evaluate plan conformance in terms of sustainability, however by advancing an evaluation framework that identifies the locations where the plan was implemented as intended, it serves as a necessary prior step to undertake such research. We must firstly be able to identify the locations where the planning intensification occurred if we are to undertake subsequent evaluations to determine if there is a corresponding improvement in measures of sustainability.

8.2.4. The prospects of achieving sustainable urban outcomes through activity centre policy Section 2.1 reviewed current literature on the purported sustainability benefits of more compact city forms. Existing research suggests such forms are correlated with changes in travel behaviour (away from automobile use), a reduction in greenhouse gas emissions (in the form of energy efficiencies), reduced traffic fatalities (but an increase in accidents), improved public health, reduced infrastructure costs, and improved economies in downtown areas. The compact activity centres policy itself is justified by appeals to such sustainability benefits. Where conformance to centre policy was low, it stands to reason that these centres would have contributed little to any sustainability gains as intended by the policy; obviously if little changed in the fundamental structure and/or intensity of land use, then any improvements or reductions in levels of sustainability in these locations must have been the result of other factors. Plan implementation is therefore a critical component in meeting the core justifications for undertaking the plan. As Gleeson (2012) argues, the difficulty of

244 implementing the compact city, and the slow incremental changes that the planning system is capable of delivering, means that the underlying justification will almost certainly fail to be met within the urgent timeframes required to avert key sustainability challenges such as climate change.

The previous discussion highlighted that if compact city policy is to be implemented, then fewer centres should be nominated and that the nomination of centres needs to have closer regard to the existing nature of the centres to ensure there is reasonable likely of development proceeding. So, what then happens to car dependent outer centres which lack the requisite socio-economic conditions for development, such as most of the centres in Logan? Are such locations destined to simply miss out on the potential sustainability benefits enjoyed by more advantaged locations? These are the questions that policy makers must consider when developing future plans and determining where to allocate resources to support future centre implementation.

Ultimately, they are facing a difficult task. The results from Chapter 5 show just how resilient the urban form proved to be to substantive change. Table 70 (p 295) compares the compactness scores for each centre in 1996 and 2016. Aside from a few notable exceptions (such as Chermside), centres that started more compact, tended to remain more compact. A good example for comparison is the mixed use score, which describes the physical structure, layout, and form of each centre. These scores remained remarkably consistent over time. Even in centres that showed reasonable conformance to policy, typically less than 15% of sites actually changed use (section 6.3). This suggests that the centres would need to undergo large scale transformations in order to see more fundamental physical changes. Without such shifts, to what extent are the changes observed in this research actually contributing to urban sustainability? Do larger hospitals, big box shopping centres, or offices in the ground floor of apartment buildings actually lead to the behavioural changes necessary to make an area more sustainable? In short, do compact activity centres contribute to urban sustainability?

As changes in transportation behaviour are responsible for many of purported sustainability benefits of compact cities (Table 1, p. 26) this area would offer a useful starting point for such research. Modelling of the employment decentralisation policies associated with activity centres in greater Brisbane does reveal some potential reductions in private vehicle use, however this assumes an idealised scenario of decentralisation of higher order employment uses tightly concentrated around the public transport nodes of centres with good public transport links (Burke et al., 2011). As previously discussed however, implementing such a scenario is easier said than done as not all centres currently have such facilities or have the realistic possibility of providing for concentrations of employment of this type. Such models

245 were the key basis for justifying the original centres concept (The State of Queensland, 1993b, 1994a), which of course assumed that the policy would be implemented. Other research does show correlations between higher density employment clusters and reduced private transport mode share for journeys to work (Loader, 2018). The results for the selected centres from greater Brisbane however looked less promising, with only marginal changes in transport mode share. Chermside for example, had the greatest levels of intensification within the 20 year study period, and was ranked as the second most compact activity centre in 2016. The changes in mode share however were less spectacular, with only a 3 percentage point change to public transport journeys to work, and a negligible change to active transport modes (Table 51).

Table 51 - Journey to work mode share in Chermside, 1996 and 2016

40 Mode share Method of 1996 2016 Travel Public 4.6% 7.8% Transport Motor 89.5% 86.7% vehicle Active 4.3% 4.4% Transport Worked 1.6% 1.1% from home

The idea of denser centres resulting in more residents being able to walk/cycle to work does not appear to have been realised. Of course, this data only captures work journeys. There may well be significant mode changes in the number trips for shopping or recreation purposes, as well as improvements in the non-transport related aspects such as efficiencies in service delivery, lower infrastructure costs, and improved recreation opportunities. These are the types of questions that should next be addressed in the evaluation of compact activity centre policy.

8.3. Directions for future research

The methods advanced by this research provide a way to evaluate compact activity centre policy implementation. Now that it is known which centres have intensified as intended, research that examines these aspects and other potential effects of the centres policy is now necessary to complete the evaluation. Based on the previous discussion, further research that considers the links between activity centre conformance and urban sustainability is an obvious

40 Source: ABS – 1996 and 2016 Census of Population and Housing, Journey to Work Data. 1996 data adjusted by inflation factor of 1.086 (see Terrill, et al. (2018, p. 60))

246

line of future inquiry. Unfortunately, the slow progress observed in greater Brisbane’s activity centres suggests the policy may not be as effective as hoped. Subsequent sustainability based evaluations are therefore essential to determine if it is worth allocating the resources necessary to persist with this approach, or if these resources could be better utilised on alternatives.

The research also suggests a number of other future lines of inquiry that emerge upon consideration of the research framework and its methods. This research has used an evaluative framework that assesses the implementation of SEQRP’s activity centre policy in terms of both conformance and performance. As discussed in detail in section 3, the difference between these approaches relates to whether the outcomes (conformance) or use (performance) of the plan is evaluated. Conformance based evaluation takes a positivist approach to planning with the view that plans are created to have effect, and that these effects should be measured to determine if the plan is achieving its intended results. Performance based evaluation however considers plans as decision making devices rather than “blueprints” for a future end state, and therefore evaluates implementation in terms of how well the plan was used when making decisions on relevant matters. However, there is clearly an inter-relationship between these concepts.

8.3.1. Directions for performance based approaches Most normative planning processes typically include a phase of evaluation in order to improve future plan quality and implementation, as well as to show the effectiveness of planning (Guyadeen & Seasons, 2018). In this research, evaluation considered using both performance and conformance perspectives proved to be a useful approach to identify deficiencies in the plan and ways to improve future plans. Without the performance component, it would not have been possible to differentiate between whether conformance issues were due to deficiencies with the plan itself and how the plan was used by key actors charged with its implementation. When undertaking this analysis of plan performance, identifying and interviewing/surveying the key actors involved in long past regulatory decisions was not considered to be practical due to the 20 year timeframe and the large number of local governments involved (several of which have been merged into other authorities and no longer exist). Performance in this research was therefore measured indirectly. This involved a deductive approach that sought evidence of performance in subordinate land use plans. Where evidence was found in terms of direct references to activity centre policy and/or policy changes reflective of the planned intent, it was deduced that decisions had been authorised to change these policies whilst having regard to the higher order regional activity centre policy. As changes to local government land use regulations were a key implementation mechanism of the regional plan, complimentary changes supported by direct acknowledgements of the

247 regional plan within key documents, were considered as evidence of performance success; i.e. there is evidence that local governments are using the regional plan policies to inform decisions on their own regulatory plans in a substantive capacity. This approach was sufficient for the evaluation undertaken in this particular research because the regional level policy was voluntarily entered into by local governments in the first ten years of the policy, and then the policy legally required local governments to use it for the second ten. It was therefore likely that local governments were going to invoke the regional policy when making their own plans, and that evidence of this would be uncovered. However, there are two key areas of future inquiry where the deductive approach to performance evaluation used here would be less appropriate.

Firstly, research that seeks to understand the reasons why particular decisions were made (or not made) would require direct engagement with key actors through either surveys, interviews, or behavioural/discursive observations. The key findings of this research have shown that the regulatory land use planning system alone is not sufficient to drive compact centre development. The documentary evidence showed that policy makers were aware of this situation from the beginning and that even after two decades of poor policy conformance, the same methods of relaxing development controls continue to be applied in the hope of a different result. Understanding why policy makers insist on this approach and how they consider plan implementation when creating and adopting policy is an important direction for future plan performance based evaluatory research. Both Bunker (2012) and Searle (2007) (and also (Bunker & Searle, 2009)) describe the characteristics of a particular Australian planning paradigm which uses metropolitan plans based on long-term population forecasts to establish development targets, which are then implemented through zoning controls. Bunker (2012, 2014) believes the continuation of this approach is a form of path dependency. Steele (2009, p. 200) identified how planners operate in “hybrid” roles which attempt to balance private and public interests with the “…task of steering competitive market aspirations to work for the public interest”. Similarly, planners’ desires to increase democratic involvement in planning have been tempered by neoliberal influenced “new public management”, creating tensions between planning ideology and neoliberalised public institutions (Sager, 2009). Perhaps planners themselves are highly doubtful of the ability of existing mechanisms to deliver compact activity centre policies but are required to rely on these mechanisms due to the institutional or political necessities of their workplaces? Or is it that neoliberal ideology has become so hegemonic that it is the “hinge” (as outlined by Pritchard (2002)) in how implementation is conceptualised by planners and other options are therefore not considered? In such a situation, planners’ expectations of what they believe can be achieved “…may be seen to exhibit certain material and social path dependencies (lock-in or irreversibility)

248

becoming the basis for future envisioning, a predisciplining of the imagination through the legacy of former expectations” (Borup et al., 2006, p. 293). Others have found that broad based regional concepts such as polycentricism are taken on as “spatial imaginaries”, enabling ambiguities which allow the justification of potentially contradictory planning decisions (Granqvist et al., 2019). The role of such institutional and sociological considerations in plan implementation is currently poorly understood and lacks empirical evidence. Research evaluating plan performance from such perspectives would offer new insights into how plans are used when implementing policy.

The second key area that would benefit from a different approach to performance evaluation relates to the examination of decisions that occurred outside of centre locations, but which potentially have an impact on centre policy. Decisions in these locations are far more likely to involve situations where centre policy is not invoked, or invoked and then not followed. Such situations may not leave documentary evidence of whether the centre policy was used or not, and would therefore require methods using more direct engagement with the actors involved in order to measure plan performance. As discussed in section 2.2, an initial centres policy position was to use the land use planning system to require restrictions on “out of centre” development (The State of Queensland, 1993b). However, a subsequent report sought a more flexible approach whereby such restrictions were only applied in the final phases of centre development (The State of Queensland, 1994a). The end result was the inclusion of a principle in the plans to limit out of centre development absent definitions of the scale and form of development to which this would apply. Prior to the commencement of regional planning in South East Queensland, local governments were already including some form of centre hierarchy in their planning regulations. Some of these centres were nominated as regional activity centres, and while local government plans changed to acknowledge the new status of these centres, they also maintained their own centre planning for the non-regional centres in their hierarchies. A more direct performance based evaluation of plan implementation would permit the investigation of land use and plan changes in other centres to determine how (or if) the regional activity centre policy was being used to inform these “out of centre” planning decisions. An expansion of the research in this capacity would help to address questions of whether more or less stringent controls on out of centre development influence centre intensification.

8.3.2. Directions for conformance based approaches Performance based assessments of plan implementation are clearly of importance in any evaluation framework. However, measuring performance absent measures of conformance, lacks real work context and begs the question, “so what?”. Alexander (2016) notes that

249 theoretical discussions around definitions of planning have become irrelevant to the concerns of actual practice where spatial planners apply their expertise to intervene in land-property markets to effect public benefit. For those actively engaged in the practice of planning, plans are effectuating devices intended to have real world impacts. The strategic planner works with communities and stakeholders to establish plans and visions for a future that offers some improvement over the status quo; places that are more efficient, more sustainable, or more interesting. The statutory planner devises regulations to align future development in the image of these neighbourhood and strategic planning goals. The development assessment planner then checks proposals for physical changes to land, structures, and use against planning regulations. This is the practical work involved in the operation of the Queensland land use planning system. The language used in plans at all levels is replete with verb led, actionable statements, clearly intended to influence physical outcomes. As the opening statement of the Ipswich Regional Centre Strategy declares in 36 point font on its own dedicated page: “Vision without action is merely a dream. Action without vision just passes the time. Vision with action can change the world.” (Barker cited in Urbis, 2008). As much as planning theorists may insist that plans are not blueprints, the review of planning documents undertaken in this research at both regional (section 2.2 ) and regulatory levels (chapter 6) indicates that these plans have a clear purpose and intent to effect real change in the physical world. As this research has demonstrated, these intentions are not always realised. Regardless of the achievement of the planning objectives however, the intent remains and in the case of South East Queensland’s activity centres this intent has been relatively consistent for 20 years. Although there are difficulties in establishing clear cause and effect between plans and outcomes, it is still reasonable and necessary to assess the basic question of whether planned objectives were achieved (section 3). For these reasons, measuring plan conformance alongside performance is considered essential to the evaluation of land use planning.

The conformance based evaluation in this research considered activity centre policy in terms of two key aspects. Firstly, the nominated activity centres were considered in terms of their compactness, using a variety of indicators, to determine to what degree the centres changed from the start of the policy in 1996 until 2016, and whether these changes conformed to the planned intent for more compact centres. The results of this evaluation enabled a subsequent analysis which compared the degree of conformance to a range of factors that potentially explained centre development. The second conformance based evaluation was more closely linked to the analysis of performance, where land use changes were compared to land use regulations in order to determine if these changes conformed to the regulatory intent. As each

250 of these land use changes involved a local government decision to approve the development41, conformance with the regulations suggests that the local government is using the regulations as intended, and they are therefore also performing. The results from the first analysis showed that the majority of centres failed to become more compact than the broader conurbation. The results of the second conformance evaluation, combined with the evaluation of performance, made it possible to conclude that the failure of the plan to meets its compactness objectives was not the result of a failure to use the plan.

The conformance based research involved the first detailed examination of land use and demographic changes in all of greater Brisbane’s activity centres. Greater Brisbane however only had a total of 21 nominated centres, which limited the available analytical methods. As most Australian capital cities make use of similar plans for activity centres, future research that includes activity centres from other cities would enable the application of more advanced approaches that would enable statistically powerful inferences to be drawn to other locations. When determining ways to measure conformance, it was not possible to simply apply previously determined methods to measure city compactness, such as those developed by Ewing and Hamidi (2014). This is not the fault of their indicators, but because the specialised nature of compact activity centre policy requires a customised set of indicators to align with the planning objectives described in the policy. As discussed in section 4.2, a range of suitable indicators were instead drawn from the literature, and selected on the basis of their alignment to policy and the availability of suitable data. These indicators were then combined into an overall score by simply averaging their z-scores. This approach has a number of precedents (Burton, 2002; Galster, et al., 2001; Stathakis & Tsilimigkas, 2014) however its limitation is that it considers all of the indicators as being of equal importance. With a larger number of centres, methods such as principal component or factor analysis could be used to account for inter-relationships between indicators and identify a smaller set of indicators that best describe the variations in the data. Similarly, the small number of centres also limits the use of statistical modelling techniques that better disentangle the cause and effect associated with the analysis of multiple, inter-related factors (as discussed in section 4.2.4 and Chapter 7). The inclusion of Sydney and Melbourne’s activity centres would add more than 100 additional centres and enable some of these techniques. However, this would require the careful analysis of the supporting policy documents to determine if the selected conformance indicators remain appropriate across jurisdictions. Similar to the previous discussion on the future possibilities for performance based evaluation, expanding the study area to include non- regionally nominated activity centres (i.e. centres identified by local governments but which

41 or to permit “self-assessment” against a set of pre-determined rules decided by the local government

251 are not included in the regional centres policy), would also be beneficial. Such an expansion would enable further testing of the findings of this research through the comparison between changes in regionally nominated centres and the non-regional centres identified by local governments independently of regional policy. Additionally, as centre locations are typically highly permissive of a wide range of possible uses, the more restrictive development controls in non-centre areas may display differences in the degree of conformance to actual land use change. Such findings could contribute further to explanations of the effects of regional centre policy on broader patterns urban development.

The main aspect inhibiting such research is the lack of suitable land use data across multiple jurisdictions. This issue was discussed in detail in section 4.2.2, and required the development of a new method that made use of observations from Google Street View and historic aerial imagery to generate a detailed land use database suitable for longitudinal analysis. The key benefit of this approach is that all the necessary data sources are currently freely available, exist in the public domain, are consistent across jurisdictions, and enable analysis at a variety of scales including the customised extents necessary to evaluate planning processes applicable to discrete geographies. Although effective, the main limitation of the application of this method to larger urban expanses is the time required to undertake the necessary observations. The task itself is not complicated, however gathering data in this manner for entire metropolitan areas could only realistically be achieved if many coders could be assembled for the task. Automated methods offer some potential, however the accuracy of these methods is currently insufficient to account for the relatively small amount of land use change observed (Li, et al., 2017). Recent government open-data initiatives are commendable, and the expansion of these policies to include more detailed land use data sets would resolve this issue. The release of local government rates databases for example, would provide point based land use data that would significantly reduce the number of required manual land use observations and prove highly useful to future research. As most Australian conurbations consist of many local governments, state government coordination would be required to ensure sufficient coverage across metropolitan extents. Unfortunately there is also a trend to commercialise urban datasets, or to require significant fees for its release. This was certainly the experience when trying to access data as part of this research, where data requests were typically refused or offered only with payment. Government property valuation data for example, has long been commercialised and recent decisions to privatise government land registry offices entirely does not bode well for this data becoming more open in the future (Abelson & Chung, 2005; Han, 2017; Willingham, 2018). There is also a potential that the use of Google Street View data may be subject to more restrictive terms of use in the future (Bader, et al., 2017), and there now exists a wide range of private companies offering to sell all manner of

252 proprietary urban data sets. The commodification of information in this manner creates barriers to future research and entrenches an exclusive system where the primary purpose of information collection is to generate profits for those with the means to afford it, rather than to resolve key urban issues in the interest of public benefit.

When expanding the study to wider areas and additional centres, it is important that the use of statistically based analytical methods do not obscure the examination of some of the intricacies of urban change. In Chapter 5, it was common to see changes to centres based on standard indicators such as density or employment, occurring due to a range of different factors. For example, some centres saw systemic changes in land uses (e.g. Chermside) across the centre extent. Change in other centres however, were often the result of change occurring on only a small number of sites (Carindale). The use of multiple categories of compactness indicators helped to highlight some of these differences when creating the centre change typology. Being able to identify these differences in development patterns is important to future analysis and could address matters such as whether a few large developments, or many smaller developments better contribute to the sustainability aims of centres policy.

8.4. A final summary

In this evaluation of greater Brisbane’s activity centre policies, a review of planning policy firstly established that the planned outcomes had remained generally consistent for a twenty year period, with on-going planning objectives to concentrate residential and employment uses within centres (section 2.2). Chapter 5 examined whether land use change conformed to plan intentions, and demonstrated that despite the conformance of some centres, the majority of centres failed to become materially more compact. Chapter 6 then considered the performance of the plan to determine if the lacklustre conformance was the result of the plan not being used (a conformance and performance failure), or if the plan was being used as intended but failed to adequately provide for its material implementation. The results demonstrated that the situation could best be explained by the latter, with Chapter 7 providing further confirmation that the centres policy failed to include an adequate implementation mechanisms, particularly for centres lacking the conditions necessary to support private sector development.

Yet if private sector development is intended to deliver activity centre outcomes, the market conditions that make this future development profitable must also exist. As the initial selection of centres did not include detailed market feasibility testing, a large number of centres were included in the centre hierarchy that lacked the necessary conditions to attract private sector

253 investment. State and local governments have made additional attempts to implement the centre policy to “make the market” through further regulatory concessions, or by directly developing centre locations. The latter of these approaches (best illustrated by the example of Ipswich city centre) showed some promise in delivering new employment generating uses. However, to date, these actions have failed to stimulate further private sector investment in higher density residential developments. Future activity centre planning therefore needs to carefully consider the intended implementation mechanisms to ensure the suitable conditions or resources exist to realise the planned objectives. If government led investments in infrastructure, services, and development are proposed, then issues of need and resources will require close investigation. If private led development is to be the principal method of implementation, planning for activity centres would require more careful consideration of the current and projected market conditions for the proposed centres’ development as part of the planning process.

But what then of the outer centres that lack adequate market conditions for development? The overall purpose of the activity centre policies was to contribute to improved sustainability outcomes. If a market led approach is pursued, are these less feasible locations destined to miss out on these benefits? Can sufficient resources be garnered for more direct approaches to be employed in support of centre development in the outer centres? Such questions, and the results of this research, raise serious doubts around the capabilities of current Australian land use planning to achieve meaningful improvements to sustainability, and highlight the need for further empirical investigation of plan implementation.

Although there is a wide field of research that links sustainability benefits with more compact urban forms, there is little research that links activity centre compactness as proposed in Australian planning policies to these sustainability benefits. This imperative highlights an obvious direction for future conformance based research that seeks relationships between policy implementation and sustainability outcomes. However, before undertaking such research it is necessary to know where past policies have been implemented and where they have not. This research has accomplished this initial task for the case of greater Brisbane and provided a range of new methods that enable similar evaluations of activity centres to be undertaken in other cities. The addition of more cases would also enable further forms of analysis that can better establish matters of causation. Such conformance based studies would of course be of interest to those who intend to see plans delivered, however there is also cause to undertake more qualitative forms of research that consider plan performance evaluation. How key policy actors make use of plans and why they make particular decisions is currently poorly understood. Such research would add sociological and institutional perspectives to

254 explanations of plan implementation which cannot otherwise be observed through the physical changes to the built environment. Regardless of the future approach however, after two decades of policy attempts, sufficient time has now passed to further evaluate whether these land use planning interventions in the urban form are yielding the promised results. This research has demonstrated that activity centre policy in greater Brisbane probably is not, and that the policy has proven difficult to implement just as predicted by early compact city critics (Birrell, et al., 2005; Breheny, 1997; Troy, 1996; Williams, 1999). As warnings of catastrophic climate change become increasingly dire (IPCC, 2018), the consequences failing to respond to sustainability issues are more pressing than ever. Now is the time for further planning evaluation to determine exactly what role planning can best play in improving urban sustainability.

255

9. References

Abadie, A., Athey, S., Imbens, G., & Wooldrige, J. (2014). Finite Population Causal Standard Errors. National Bureau of Economic Research Working Papers, Paper 20325. Retrieved from: https://www.nber.org/papers/w20325.pdf

Abbott, J. (1995). Seq 2001: Quality Strategic Planning for South East Queensland. Australian Planner, 32(3), 135-138. doi:10.1080/07293682.1995.9657675

Abbott, J. (2001). A Partnership Approach to Regional Planning in South East Queensland. Australian Planner, 38(3-4), 114-120. doi:10.1080/07293682.2001.9657955

Abdullahi, S., Pradhan, B., & Jebur, M. N. (2014). GIS-based sustainable city compactness assessment using integration of MCDM, Bayes theorem and RADAR technology. Geocarto International, 30(4), 365-387. doi:10.1080/10106049.2014.911967

Abelson, P., & Chung, D. (2005). The real story of housing prices in Australia from 1970 to 2003. Australian Economic Review, 38(3), 265-281. doi:https://doi.org/10.1111/j.1467- 8462.2005.00373.x

ABS. (1996). 2033.0.55.001 - Census of Population and Housing: Socio-Economic Indexes for Areas (SEIFA), Australia - Data only, 1996. http://www.abs.gov.au/AUSSTATS/[email protected]/allprimarymainfeatures/F41FEC0CFE5EF0EECA 25715B007A7D4D?opendocument

ABS, (1998). Information Paper: Socio-Economic Indexes For Areas 2039.0. Retrieved from http://www.ausstats.abs.gov.au/ausstats/free.nsf/0/C17E9A880591BB45CA256AE9001BCD 57/$File/2039.0_1996.pdf

ABS, (2008). Socio-economic indexes for areas (SEIFA) - Technical paper 2039.0.55.001. Retrieved from http://www.ausstats.abs.gov.au/ausstats/subscriber.nsf/0/72283F45CB86E5FECA25741700 11B271/$File/2039055001_socio-economic%20indexes%20for%20areas%20(seifa)%20- %20technical%20paper_2006.pdf

ABS. (2016). 2901.0 - Census of Population and Housing: Census Dictionary, 2016. http://www.abs.gov.au/ausstats/[email protected]/Lookup/2901.0Chapter38202016

ABS (2017, June). [Personal communication with authors].

256

Albrechts, L., & Balducci, A. (2013). Practicing Strategic Planning: In Search of Critical Features to Explain the Strategic Character of Plans. disP - The Planning Review, 49(3), 16-27. doi:10.1080/02513625.2013.859001

Alexander, E. R. (2006). Evolution and Status: Where is Planning-Evaluation Today and How Did it Get Here? In E. R. Alexander (Ed.), Evaluation in Planning: Evolution and Prospects (pp. 3-16). Hampshire, UK: Ashgate.

Alexander, E. R. (2016). There is no planning—only planning practices: Notes for spatial planning theories. Planning Theory, 15(1), 91-103. doi:10.1177/1473095215594617

Alexander, E. R., & Faludi, A. (1989). Planning and plan implementation: notes on evaluation criteria. Environment and Planning B: Planning and Design, 16(2), 127-140. doi:10.1068/b160127

Alfasi, N., Almagor, J., & Benenson, I. (2012). The actual impact of comprehensive land-use plans: Insights from high resolution observations. Land Use Policy, 29(4), 862-877. doi:10.1016/j.landusepol.2012.01.003

Allen, M., & Yen, W. (2001). Introduction to measurement theory: Waveland Press.

Alterman, R., & Hill, M. (1978). Implementation of Urban Land Use Plans. Journal of the American Institute of Planners, 44(3), 274-285. doi:10.1080/01944367808976905

Altes, W. K. K. (2006). Stagnation in Housing Production: Another Success in the Dutch ‘Planner's Paradise’? Environment and Planning B: Planning and Design, 33(1), 97-114. doi:10.1068/b31192

Andrews, C. J. (2008). Greenhouse gas emissions along the rural-urban gradient. Journal of Environmental Planning and Management, 51(6), 847-870. doi:10.1080/09640560802423780

Anthony, R. N. (1965). Planning and Control Systems: A Framework for Analysis: Division of Research, Graduate School of Business Administration, Harvard University.

Aurand, A. (2013). Does Sprawl Induce Affordable Housing? Growth and Change, 44(4), 631-649. doi:10.1111/grow.12024

AURIN. (2016). Neighbourhood Generator. https://docs.aurin.org.au/portal-help/analysing-your- data/walkability-tools/walkability001-points-to-regions/

Bader, M. D. M., Mooney, S. J., Bennett, B., & Rundle, A. G. (2017). The Promise, Practicalities, and Perils of Virtually Auditing Neighborhoods Using Google Street View. The ANNALS of the

257

American Academy of Political and Social Science, 669(1), 18-40. doi:10.1177/0002716216681488

Berke, P., Backhurst, M., Day, M., Ericksen, N., Laurian, L., Crawford, J., & Dixon, J. (2006). What Makes Plan Implementation Successful? An Evaluation of Local Plans and Implementation Practices in New Zealand. Environment and Planning B: Planning and Design, 33(4), 581-600. doi:10.1068/b31166

Birrell, B., O’Connor, K., Rapson, V., & Healy, E. (2005). Melbourne 2030: Planning rhetoric versus urban reality. Melbourne: Monash University Press.

BITRE. (2013). Population growth, jobs growth and commuting flows in South East Queensland, Report 134. Canberra, ACT. Retrieved from http://bitre.gov.au/publications/2013/report_134.aspx

Borup, M., Brown, N., Konrad, K., & Van Lente, H. (2006). The sociology of expectations in science and technology. Technology Analysis & Strategic Management, 18(3-4), 285-298. doi:10.1080/09537320600777002

Breheny, M. (1995). The Compact City and Transport Energy Consumption. Transactions of the Institute of British Geographers, 20(1), 81-101. doi:10.2307/622726

Breheny, M. (1997). Urban compaction: feasible and acceptable? Cities, 14(4), 209-217. doi:10.1016/S0264-2751(97)00005-X

Brennan, R. L., & Prediger, D. J. (2016). Coefficient Kappa: Some Uses, Misuses, and Alternatives. Educational and Psychological Measurement, 41(3), 687-699. doi:10.1177/001316448104100307

Brewer, K., & Grant, J. L. (2015). Seeking density and mix in the suburbs: challenges for mid-sized cities. Planning Theory and Practice, 16(2), 151-168. doi:10.1080/14649357.2015.1011216

Brody, S. D., & Highfield, W. E. (2005). Does Planning Work?: Testing the Implementation of Local Environmental Planning in Florida. Journal of the American Planning Association, 71(2), 159- 175. doi:10.1080/01944360508976690

Brody, S. D., Highfield, W. E., & Thornton, S. (2006). Planning at the Urban Fringe: An Examination of the Factors Influencing Nonconforming Development Patterns in Southern Florida. Environment and Planning B: Planning and Design, 33(1), 75-96. doi:10.1068/b31093

Brueckner, J., & Rosenthal, S. (2009). Gentrification and Neighborhood Housing Cycles: Will America's Future Downtowns be Rich? The Review of Economics and Statistics, 91(4), 725-743. Retrieved from: https://www.mitpressjournals.org/doi/pdfplus/10.1162/rest.91.4.725

258

Bryant, L. (2013). Urban consolidation - Reality or fantasy? Planning News, 39(7), 28-29. Retrieved from: http://search.informit.com.au/documentSummary;dn=473857648082160;res=IELBUS

Building Owners and Managers Association Queensland Division. (1993). Directory of Shopping Centres. Retrieved from Trove

Bujang, M., & Baharum, N. (2017). Guidelines of the minimum sample size requirements for Cohen’s Kappa. Epidemiology Biostatistics and Public Health, 14(2). doi:10.2427/12267

Bunker, R. (2012). Reviewing the Path Dependency in Australian Metropolitan Planning. Urban Policy and Research, 30(4), 443-452. doi:10.1080/08111146.2012.700638

Bunker, R. (2014). How Is the Compact City Faring in Australia? Planning Practice & Research, 29(5), 449-460. doi:10.1080/02697459.2014.945376

Bunker, R., & Searle, G. (2009). Theory and Practice in Metropolitan Strategy: Situating Recent Australian Planning. Urban Policy and Research, 27(2), 101-116. doi:10.1080/08111140902971426

Burby, R. J. (2003). Making Plans that Matter:Citizen Involvement and Government Action. Journal of the American Planning Association, 69(1), 33-49. doi:10.1080/01944360308976292

Burke, M. I., Li, T., & Dodson, J. (2011). What Happens When Government Workers Move to the Suburbs? Transportation Research Record: Journal of the Transportation Research Board, 2255(1), 110-116. doi:10.3141/2255-12

Burton, E. (2002). Measuring Urban Compactness in UK Towns and Cities. Environment and Planning B: Planning and Design, 29(2), 219-250. doi:10.1068/b2713

Buxton, M., & Tieman, G. (2005). Patterns of Urban Consolidation in Melbourne: Planning Policy and the Growth of Medium Density Housing. Urban Policy and Research, 23(2), 137-157. doi:10.1080/08111470500135086

Canepa, B. (2007). Bursting the Bubble: Determining the Transit-Oriented Development's Walkable Limits. Transportation Research Record: Journal of the Transportation Research Board, 1992, 28-34. doi:10.3141/1992-04

Cantor, A. (1996). Sample-size calculations for Cohen's kappa. Psychological Methods, 1(2), 150-153. Retrieved from: http://dx.doi.org.ezp01.library.qut.edu.au/10.1037/1082-989X.1.2.150

Carruthers, J. I., & Ulfarsson, G. F. (2003). Urban Sprawl and the Cost of Public Services. Environment and Planning B: Planning and Design, 30(4), 503-522. doi:10.1068/b12847

259

Cervero, R. (2001). Efficient Urbanisation: Economic Performance and the Shape of the Metropolis. Urban Studies, 38(10), 1651-1671. doi:10.1080/00420980120084804

Chakraborty, A., Knaap, G.-J., Nguyen, D., & Jung Ho, S. (2009). The Effects of High-density Zoning on Multifamily Housing Construction in the Suburbs of Six US Metropolitan Areas. Urban Studies, 47(2), 437-451. doi:10.1177/0042098009348325

Chapin, T. S., Deyle, R. E., & Baker, E. J. (2008). A Parcel-Based GIS Method for Evaluating Conformance of Local Land-Use Planning with a State Mandate to Reduce Exposure to Hurricane Flooding. Environment and Planning B: Planning and Design, 35(2), 261-279. doi:10.1068/b32114

Chhetri, P., Han, J. H., Chandra, S., & Corcoran, J. (2013). Mapping urban residential density patterns: Compact city model in Melbourne, Australia. City, Culture and Society, 4(2), 77-85. doi:10.1016/j.ccs.2013.03.001

Churchman, A. (1999). Disentangling the Concept of Density. Journal of Planning Literature, 13(4), 389- 411. doi:10.1177/08854129922092478

City of Sydney, (2012a). City of Sydney Floor Space and Employment Survey.

City of Sydney, (2012b). Crown and Baptist Streets Village Summary Report: City of Sydney Floor Space and Employment Survey.

City of Sydney, (2012c). Glebe Point Road Village Summary Report: City of Sydney Floor Space and Employment Survey.

City of Sydney, (2012d). Harris Street Village Summary Report: City of Sydney Floor Space and Employment Survey.

City of Sydney, (2012e). King Street Village Summary Report: City of Sydney Floor Space and Employment Survey.

City of Sydney, (2012f). Macleay Street and Wolloomooloo Village Summary Report: City of Sydney Floor Space and Employment Survey.

City of Sydney, (2012g). Oxford Street Village Summary Report: City of Sydney Floor Space and Employment Survey.

City of Sydney, (2012h). Redfern Street Village Summary Report: City of Sydney Floor Space and Employment Survey.

260

Coffee, N. T., Lange, J., & Baker, E. (2016). Visualising 30 Years of Population Density Change in Australia’s Major Capital Cities. Australian Geographer, 47(4), 511-525. doi:10.1080/00049182.2016.1220901

Cohen, J. (1960). A Coefficient of Agreement for Nominal Scales. Educational and Psychological Measurement, 20(1), 37-46. Retrieved from: https://journals.sagepub.com/doi/pdf/10.1177/001316446002000104

Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences: Routledge.

Connell, D. J., & Daoust-Filiatrault, L.-A. (2018). Better Than Good: Three Dimensions of Plan Quality. Journal of Planning Education and Research, 38(3), 265-272. doi:10.1177/0739456x17709501

Cook, N., Taylor, E., Hurley, J., & Colic-Peisker, V. (2012). Resident third party objections and appeals against planning applications: implications for higher density and social housing. Australian Housing and Urban Research Institute, Retrieved from https://www.ahuri.edu.au/research/final-reports/197

Corder, G., & Foreman, D. (2014). Nonparametric Statistics : A Step-By-Step Approach: John Wiley & Sons.

Cromwell. (2011). Ipswich City Heart Trust: Product disclosure statement. Retrieved from https://www.cromwell.com.au/__data/assets/pdf_file/0033/18996/Ipswich-City-Heart- Product-Disclosure-Statement-Dec-2011.pdf

Curtis, C., & Scheurer, J. (2015). Performance measures for public transport accessibility: Learning from international practice. Journal of Transport and Land Use. doi:10.5198/jtlu.2015.683

Curtis, J. W., Curtis, A., Mapes, J., Szell, A. B., & Cinderich, A. (2013). Using Google Street View for systematic observation of the built environment: analysis of spatio-temporal instability of imagery dates. International Journal of Health Geographics, 12, 53. doi:10.1186/1476-072X- 12-53

Dain, A. (2005). Residential Land-Use Regulation in Eastern Massachusetts: A Study of 187 Communities. Rappaport Institute for Greater Boston Harvard University and Boston: Pioneer Institute for Public Policy Research, Cambridge. Retrieved from http://www.masshousingregulations.com/pdf/land_use_regulation.pdf

Davidson, G. (1997). The Great Australian Sprawl. Historic Environment, 13(1), 10-17. Retrieved from: https://search.informit.com.au/documentSummary;dn=769402465333798;res=IELHSS

261

Day, J., Starup, S., Chen, Y., Budahazy, M., Wu, A., & Fan, L. (2015). Activity Centre Policy Effects on Employment Clustering: A Spatial Study of Job Density in Melbourne, Australia. Paper presented at State of Australian Cities Conference, Gold Coast, Queensland.

Day, J., Sturup, S., Chen, Y., & Budahazy, M. (2014). An open-source tool for identifying industry clusters: a demonstration in the Northwest Corridor of Melbourne, Australia. In Australian Regional Development Conference.

de Vries, H. (1993). The rowwise correlation between two proximity matrices and the partial rowwise correlation. Psychometrika, 58(1), 53-69. doi:10.1007/bf02294470

Dey, C., Berger, C., Foran, B., Foran, M., Joske, R., Lenzen, M., & Wood, R. (2007). Household Environmental Pressure from Consumption: An Australian Environmental Atlas. In G. Birch (Ed.), Water, Wind, Art and Debate: How Environmental Concerns Impact on Disciplinary Research (pp. 280-315). Sydney: University of Sydney Press.

Dodson, J. (2010). In the Wrong Place at the Wrong Time? Assessing some Planning, Transport and Housing Market Limits to Urban Consolidation Policies. Urban Policy and Research, 28(4), 487- 504. doi:10.1080/08111146.2010.517158

Dodson, J. (2012). Transforming Australia's 'housing solution': How We Can Better Plan Suburbia to Meet Our Future Challenges. In R. Tomlinson (Ed.), Australia's Unintended Cities: The Impact of Housing on Urban Development (pp. 19-34). Victoria, Australia: CSIRO Publishing.

Echenique, M. H., Hargreaves, A. J., Mitchell, G., & Namdeo, A. (2012). Growing Cities Sustainably. Journal of the American Planning Association, 78(2), 121-137. doi:10.1080/01944363.2012.666731

Evans, J. (1996). Straightforward Statistics for the Behavioral Sciences: Sage.

Evenson, B., & Wheaton, W. (2003). Local Variation in Land Use Regulations. Brrokings-Wharton Papers on Urban Affairs, 221-260. Retrieved from: https://www.jstor.org/stable/pdf/25067399.pdf

Ewing, R., & Cervero, R. (2010). Travel and the Built Environment. Journal of the American Planning Association, 76(3), 265-294. doi:10.1080/01944361003766766

Ewing, R., & Hamidi, S. (2014). Measuring urban sprawl and validating sprawl measures. National Institutes of Health and Smart Growth Amercia, Washington DC. Retrieved from https://gis.cancer.gov/tools/urban-sprawl/sprawl-report-short.pdf

Ewing, R., & Hamidi, S. (2015). Compactness versus Sprawl: A Review of Recent Evidence from the United States. Journal of Planning Literature, 30(4), 413-432. doi:10.1177/0885412215595439

262

Ewing, R., Meakins, G., Hamidi, S., & Nelson, A. C. (2014). Relationship between urban sprawl and physical activity, obesity, and morbidity – Update and refinement. Health & Place, 26, 118- 126. doi:http://dx.doi.org/10.1016/j.healthplace.2013.12.008

Ewing, R., Pendall, R., & Chen, D. (2002). Measuring Sprawl and its Impact. Smart Growth America, Retrieved from https://www.smartgrowthamerica.org/app/legacy/documents/MeasuringSprawl.PDF

Ewing, R., & Rong, F. (2008). The impact of urban form on U.S. residential energy use. Housing Policy Debate, 19(1), 1-30. doi:10.1080/10511482.2008.9521624

Faludi, A. (2000). The Performance of Spatial Planning. Planning Practice and Research, 15(4), 299- 318. doi:10.1080/713691907

Faludi, A. (2006). Evaluating Plans: The Application of the European Spatial Development Perspective. In E. R. Alexander (Ed.), Evaluation in Planning: Evolution and Prospects (pp. 119-143). Hampshire, UK: Ashgate.

Faludi, A., & Altes, W. K. (1994). Evaluating communicative planning: A revised design for performance research. European Planning Studies, 2(4), 403-418. doi:10.1080/09654319408720278

Feitelson, E., Felsenstein, D., Razin, E., & Stern, E. (2017). Assessing land use plan implementation: Bridging the performance-conformance divide. Land Use Policy, 61, 251-264. doi:10.1016/j.landusepol.2016.11.017

Filion, P. (2009). The mixed success of nodes as a smart growth planning policy. Environment and Planning B: Planning and Design, 36(3), 505-521. doi:10.1068/b33145

Fleiss, J. (1981). Statistical methods for rates and proportions. New York: Wiley.

Forster, C. (2004). Australian Cities: Continuity and Change (3rd ed.). Melbourne, Australia: Oxford University Press.

Forster, C. (2006). The challenge of change: Australian cities and urban planning in the new millennium. Geographical Research, 44(2), 173-182. doi:10.1111/j.1745-5871.2006.00374.x

Freestone, R. (2012). An Historical Perspective. In S. Thompson & P. J. Maginn (Eds.), Planning Australia: An Overview of Urband and Regional Planning (2nd ed., pp. 73-97). New York: Cambridge University Press.

Gabbe, C. J. (2018). Why Are Regulations Changed? A Parcel Analysis of Upzoning in Los Angeles. Journal of Planning Education and Research, 289-300. doi:10.1177/0739456x17696034

263

Galster, G., Hanson, R., Ratcliffe, M. R., Wolman, H., Coleman, S., & Freihage, J. (2001). Wrestling Sprawl to the Ground: Defining and measuring an elusive concept. Housing Policy Debate, 12(4), 681-717. doi:10.1080/10511482.2001.9521426

Gamer, M., Lemon, J., Fellows, I., & Singh, P. (2019). Package 'irr'. Retrieved from https://www.r- project.org

Gardiner, Y. (2009, 24 February). Treasury signs off on city square loan, The Queensland Times. p. 1.

Gehl, J., Kaefer, L. J., & Reigstad, S. (2006). Close encounters with buildings. Urban Design International, 11(1), 29-47. doi:10.1057/palgrave.udi.9000162

Gennaio, M.-P., Hersperger, A. M., & Bürgi, M. (2009). Containing urban sprawl—Evaluating effectiveness of urban growth boundaries set by the Swiss Land Use Plan. Land Use Policy, 26(2), 224-232. doi:10.1016/j.landusepol.2008.02.010

Giuliano, G., & Small, K. A. (1991). Subcenters in the Los Angeles region. Regional Science and Urban Economics, 21(2), 163-182. doi:10.1016/0166-0462(91)90032-i

Glaeser, E. L., Schuetz, J., & Ward, B. (2006). Regulation and the rise of housing prices in Greater Boston. Rappaport Institute for Greater Boston Harvard University and Boston: Pioneer Institute for Public Policy Research, Cambridge. Retrieved from http://www.masshousingregulations.com/index.asp

Gleeson, B. (2012). ‘Make No Little Plans’: Anatomy of Planning Ambition and Prospect. Geographical Research, 50(3), 242-255. doi:10.1111/j.1745-5871.2011.00728.x

Gleeson, B., Dodson, J., & Spiller, M. (2012). Governance, Metropolitan Planning and City-Building: the Case for Reform. In R. Tomlinson (Ed.), Australia's Unintended Cities: The Impact of Housing on Urban Development (pp. 117-134). Victoria, Australia: CSIRO Publishing.

Gleeson, B., & Low, N. (2000a). Australian Urban Planning: New Challenges, New Agendas. Sydney, Australia: Allen Unwin.

Gleeson, B., & Low, N. (2000b). ‘Unfinished Business’: Neoliberal Planning Reform in Australia. Urban Policy and Research, 18(1), 7-28. doi:10.1080/08111140008727821

Goodman, R., & Coote, M. (2007). Sustainable Urban Form and the Shopping Centre: An Investigation of Activity Centres in Melbourne's Growth Areas. Urban Policy and Research, 25(1), 39-61. doi:10.1080/08111140701225578

264

Gordon, P., & Richardson, H. W. (1996). Beyond Polycentricity: The Dispersed Metropolis, Los Angeles, 1970-1990. Journal of the American Planning Association, 62(3), 289-295. doi:10.1080/01944369608975695

Gordon, P., & Richardson, H. W. (1997). Are Compact Cities a Desirable Planning Goal? Journal of the American Planning Association, 63(1), 95-106. doi:10.1080/01944369708975727

Granqvist, K., Sarjamo, S., & Mäntysalo, R. (2019). Polycentricity as spatial imaginary: the case of Helsinki City Plan. European Planning Studies, 27(4), 739-758. doi:10.1080/09654313.2019.1569596

Grant, J. L. (2009). Theory and Practice in Planning the Suburbs: Challenges to Implementing New Urbanism, Smart Growth, and Sustainability Principles. Planning Theory & Practice, 10(1), 11- 33. doi:10.1080/14649350802661683

Griew, P., Hillsdon, M., Foster, C., Coombes, E., Jones, A., & Wilkinson, P. (2013). Developing and testing a street audit tool using Google Street View to measure environmental supportiveness for physical activity. International Journal of Behavioral Nutrition and Physical Activity, 10, 103. doi:10.1186/1479-5868-10-103

Grosvenor, M. (2013). Does Urban Form Influence Sustainable Behaviour Doctor of Philosophy. University of Western Sydney, Sydney.

Guyadeen, D., & Seasons, M. (2016). Plan Evaluation: Challenges and Directions for Future Research. Planning Practice & Research, 31(2), 215-228. doi:10.1080/02697459.2015.1081335

Guyadeen, D., & Seasons, M. (2018). Evaluation Theory and Practice: Comparing Program Evaluation and Evaluation in Planning. Journal of Planning Education and Research, 38(1), 98-110. doi:10.1177/0739456x16675930

Gyourko, J., & Molloy, R. (2015). Regulation and housing supply. In G. Duranton, J. Henderson & W. Strange (Eds.), Handbook of Regional and Urban Economics (Vol. 5, pp. 1289-1337): Elsevier.

Gyourko, J., Saiz, A., & Summers, A. (2008). A New Measure of the Local Regulatory Environment for Housing Markets: The Wharton Residential Land Use Regulatory Index. Urban Studies, 45(3), 693-729. doi:10.1177/0042098007087341

Han, E. (2017). NSW land titles registry leased for $2.6 billion to Hastings Funds Management, First State Super, The Sydney Morning Herald. Retrieved from https://www.smh.com.au/national/nsw/nsw-land-titles-registry-leased-for-26-billion-to- hastings-funds-management-first-state-super-20170412-gvjcfw.html

265

Hawley, K., & Moellering, H. (2005). A Comparative Analysis of Areal Interpolation Methods. Cartography and Geographic Information Science, 32(4), 411-423. doi:10.1559/152304005775194818

Heffernan, E., Heffernan, T., & Pan, W. (2013). The relationship between the quality of active frontages and public perceptions of public spaces. URBAN DESIGN International, 19(1), 92-102. doi:10.1057/udi.2013.16

Hinklle, D., Wiersma, W., & Jurs, S. (2003). Applied Statistics for the Behavioral Sciences. Boston: Houghton Mifflin.

Howarth, A. (2019). Closing up for the final time, The Ipswich Advertiser. p. 5.

Hussain, E., & Shan, J. (2016). Urban building extraction through object-based image classification assisted by digital surface model and zoning map. International Journal of Image and Data Fusion, 7(1), 63-82. doi:10.1080/19479832.2015.1119206

Hwang, J., & Sampson, R. J. (2014). Divergent Pathways of Gentrification: Racial Inequality and the Social Order of Renewal in Chicago Neighborhoods. American Sociological Review, 79(4), 726- 751. doi:10.1177/0003122414535774

IPCC. (2018). Global Warming of 1.5C. Retrieved from https://www.ipcc.ch/sr15/

Ipswich City Council, (2013). Development Application Negotiated Decision Notice: 2107/2012/MCU. Retrieved from http://pdonline.ipswich.qld.gov.au

Ipswich City Properties, (2009). Ipswich City Heart Project: Invitation for expressions of interest.

Ipswich City Properties, (2017). Ipswich City Centre Master Plan Presentation. Retrieved from http://www.ipswich-commercial.com.au/wp-content/uploads/2017/09/170823_Master- Plan-Presentation_MEDIA-2_Current-Sept-2017.pdf

Jackson, Z. (2010, 16 December). $1 billion plan for the heart of Ipswich, The Queensland Times. p. 2.

Jacobs, J. (1993). The Death and Life of Great American Cities. New York: Modern Library.

Jain, A., & Courvisanos, J. (2009). Urban Growth Centres on the Periphery: Ad Hoc Policy Vision and Research Negelect. Australasian Journal of Regional Studies, 15(1), 3-26. Retrieved from: https://search.proquest.com/docview/220952735

Johnson, H. (2018, 17 October). City’s CBD company will close: Council to manage mall project, The Queensland Times. p. 6.

266

Kahn, M. (2007). Quality of Life and Productivity in Sprawled Versus Compact US Cities. OECD Publishing, unpublished manuscript.

Kelly, C. M., Wilson, J. S., Baker, E. A., Miller, D. K., & Schootman, M. (2013). Using Google Street View to audit the built environment: inter-rater reliability results. Annals of Behavioral Medicine, 45(1), 108-112. doi:10.1007/s12160-012-9419-9

Kendall, M. (1962). Rank correlation methods (3rd ed.). London: Griffin.

Ker, I., & Ginn, S. (2003). Myths and realities in walkable catchments: The case of walking and transit. Road & Transport Research, 12(2), 69-80. Retrieved from: http://gateway.library.qut.edu.au/login?url=http://search.proquest.com.ezp01.library.qut.e du.au/docview/215247207?accountid=13380

Kickert, C. C. (2015). Active centers – interactive edges: The rise and fall of ground floor frontages. Urban Design International, 21(1), 55-77. doi:10.1057/udi.2015.27

Kline, J., & Alig, R. (1999). Does Land Use Planning Slow the Conversion of Fores and Farm Lands? Growth and Change, 30, 3-22. Retrieved from: https://onlinelibrary.wiley.com/doi/pdf/10.1111/0017-4815.00102

Krugman, P. (2009). How Did Economists Get it So Wrong, The New York Times Magazine. Retrieved from https://www.nytimes.com/2009/09/06/magazine/06Economic-t.html

Laurian, L., Crawford, J., Day, M., Kouwenhoven, P., Mason, G., Ericksen, N., & Beattie, L. (2010). Evaluating the Outcomes of Plans: Theory, Practice, and Methodology. Environment and Planning B: Planning and Design, 37(4), 740-757. doi:10.1068/b35051

Laurian, L., Day, M., Berke, P., Ericksen, N., Backhurst, M., Crawford, J., & Dixon, J. (2004). Evaluating Plan Implementation: A Conformance-Based Methodology. Journal of the American Planning Association, 70(4), 471-480. doi:10.1080/01944360408976395

Lavrakas, P. (2008). Encyclopedia of Survey Research Methods. Thousand Oaks, CA: Sage.

Li, X., Zhang, C., & Li, W. (2017). Building block level urban land-use information retrieval based on Google Street View images. GIScience & Remote Sensing, 1-17. doi:10.1080/15481603.2017.1338389

Lichfield, N., Kettle, P., & Whitbread, M. (1975). Evaluation in the planning process. Oxford, UK: Pergamon.

267

Limb, M., Grodach, C., & Mayere, S. (2018). Where is the compact city?: A method to monitor the implementation of compact activity centre policy. Paper presented at State of Australian Cities Conference, Adelaide, Australia.

Lin, J.-J., & Yang, A.-T. (2006). Does the Compact-City Paradigm Foster Sustainability? An Empirical Study in Taiwan. Environment and Planning B: Planning and Design, 33(3), 365-380. doi:10.1068/b31174

Lin, L. I. K. (1989). A Concordance Correlation Coefficient to Evaluate Reproducibility. Biometrics, 45(1), 255. doi:10.2307/2532051

Liu, X., Pan, Q., King, L., & Jin, Z. (2018). Analysing the changes of employment subcentres: A comparison study of Houston and Dallas. Urban Studies, 004209801878955. doi:10.1177/0042098018789554

Loader, C. (2018). Suburban employment clusters and the journey to work in Australian cities. https://chartingtransport.com/2018/07/08/suburban-employment-clusters-and-the- journey-to-work-in-australian-cities/

Logan City Council, (2009). Greater Springwood Master Plan. Retrieved from http://www.logan.qld.gov.au/__data/assets/pdf_file/0016/4723/greaterspringwood.pdf

Logan City Council. (2016). Springwood Summit. http://www.logan.qld.gov.au/planning-and- building/planning-and-development/springwood-summit

Logan City Council, Centres Planning and Urban Renewal. (2017a). Springwood Implementation Plan. Retrieved from https://www.logan.qld.gov.au/__data/assets/pdf_file/0010/439183/Springwood- Implementation-Plan-Ver4-171109.pdf

Logan City Council, Centres Planning and Urban Renewal. (2017b). Springwood Summit Outcomes. Retrieved from https://www.logan.qld.gov.au/__data/assets/pdf_file/0011/417458/Springwood-Summit- Outcomes.pdf

Loh, C. G. (2011). Assessing and Interpreting Non-conformance in Land-use Planning Implementation. Planning Practice and Research, 26(3), 271-287. doi:10.1080/02697459.2011.580111

Lonsdale, R. E. (1972). Manufacturing Decentralization: The Discouraging Record in Australia. Land Economics, 48(4), 321. doi:10.2307/3145308

268

Lyles, W., Berke, P., & Smith, G. (2015). Local plan implementation: assessing conformance and influence of local plans in the United States. Environment and Planning B: Planning and Design, 43(2), 381-400. doi:10.1177/0265813515604071

Lyles, W., & Stevens, M. (2014). Plan Quality Evaluation 1994-2012: Growth and Contributions, Limitations, and New Directions. Journal of Planning Education and Research, 34(4), 433-450. doi:10.1177/0739456x14549752

Mavoa, S., Eagleson, S., Badland, H. M., Gunn, L., Boulange, C., Stewart, J., & Giles-Corti, B. (2018). Identifying appropriate land-use mix measures for use in a national walkability index. Journal of Transport and Land Use, 11(1). doi:10.5198/jtlu.2018.1132

Maxwell, S. E. (2000). Sample size and multiple regression analysis. Psychological Methods, 5(4), 434- 458. doi:10.1037/1082-989x.5.4.434

McLoughlin, J. B. (1992). Shaping Melbourne's Future? Town Planning, the State and Civil Society. Cambridge: Cambridge University Press.

Mees, P. (2009). How dense are we? Another look at urban density and transport patterns in Australia, Canada and the USA. Road & Transport Research, 18(4), 58-67. Retrieved from: http://search.informit.com.au/documentSummary;dn=738822610697382;res=IELENG

Mennis, J. (2015). Increasing the Accuracy of Urban Population Analysis With Dasymetric Mapping. Cityscape: A Journal of Policy Development and Research, 17(1), 115-126. Retrieved from:

Morikawa, M. (2011). Economies of Density and Productivity in Service Industries: An Analysis of Personal Service Industries Based on Establishment-Level Data. Review of Economics and Statistics, 93(1), 179-192. doi:10.1162/REST_a_00065

Myers, J., Well, A., & Lorch, R. (2010). Research Design and Statistical Analysis: Routledge.

Myors, P., O'Leary, R., & Helstroom, R. (2005). Multi Unit Residential Buildings Energy & Peak Demand Study. Energy Australia, Sydney, Australia. Retrieved from http://www.ausgrid.com.au/Common/Our-network/Demand-management-and-energy- efficiency/Energy-use-facts-and- figures/~/media/Files/Network/Demand%20Management/Energy%20use%20resources/Net works_multi_unit_sumrep_Oct08.pdf

Naik, N., Kominers, S. D., Raskar, R., Glaeser, E. L., & Hidalgo, C. A. (2017, Jul 06). Computer vision uncovers predictors of physical urban change. In Proceedings of the National Academy of Sciences.

269

Nematollahi, S., Tiwari, R., & Hedgecock, D. (2016). Desirable Dense Neighbourhoods: An Environmental Psychological Approach for Understanding Community Resistance to Densification. Urban Policy and Research, 34(2), 132-151. doi:10.1080/08111146.2015.1078233

Newman, P., & Kenworthy, J. R. (1992). Winning Back the Cities. Marrickville, Australia: Australian Consumers' Association.

Newman, P., & Kenworthy, J. R. (1999). Sustainability and Cities: Overcoming Automobile Dependence. Washington D.C.: Island Press.

Newman, P. W. G., & Kenworthy, J. R. (1989). Gasoline Consumption and Cities. Journal of the American Planning Association, 55(1), 24-37. doi:10.1080/01944368908975398

Newton, P. (1997). Conclusions and Strategies. In P. Newton (Ed.), Re-Shaping Cities: Re-Shaping Cities for a More Sustainable Future - Exploring the Link Between Urban Form, Air Quality, Energy and Greenhouse Gas Emissions. Melbourne, Australia: Australian Housing and Urban Research Institute (AHURI).

Newton, P., & Glackin, S. (2014). Understanding Infill: Towards New Policy and Practice for Urban Regeneration in the Established Suburbs of Australia's Cities. Urban Policy and Research, 32(2), 121-143. doi:10.1080/08111146.2013.877389

Newton, P., Murray, S., Wakefield, R., Murphy, C., Khor, L. A., & Morgan, T. (2011). Towards a new development model for housing regeneration in greyfield residential precincts. Australian Housing and Urban Research Institute, Melbourne. Retrieved from http://www.ahuri.edu.au/research/final-reports/171

O'Connor, K. (2003). Melbourne 2030: A Response. Urban Policy and Research, 21(2), 211-215. doi:10.1080/08111140309955

O'Connor, K., & Healy, E. (2004). Rethinking suburban development in Australia: a Melbourne case study. European Planning Studies, 12(1), 27-40. doi:10.1080/09654310310001635698

OECD. (2012). Compact City Policies: A Comparative Assessment. Paris: OECD Publishing.

Olesen, K. (2014). The neoliberalisation of strategic spatial planning. Planning Theory, 13(3), 288-303. doi:10.1177/1473095213499340

Oliveira, V., & Pinho, P. (2009). Evaluating Plans, Processes and Results. Planning Theory & Practice, 10(1), 35-63. doi:10.1080/14649350802661741

270

Oliveira, V., & Pinho, P. (2010a). Evaluation in Urban Planning: Advances and Prospects. Journal of Planning Literature, 24(4), 343-361. doi:10.1177/0885412210364589

Oliveira, V., & Pinho, P. (2010b). Measuring success in planning:Developing and testing a methodology for planning evaluation. Town Planning Review, 81(3), 307-332. doi:10.3828/tpr.2010.7

Orr, S. (2016, 21 October). Investment opportunities and barriers for new CBD discussed at Springwood Summit, Quest Newspapers. Retrieved from http://www.couriermail.com.au/questnews/logan/investment-opportunities-and-barriers- for-new-cbd-discussed-at-springwood-summit/news- story/986bc41f30db9694f1d893749fdf4135

Padeiro, M. (2014). The influence of transport infrastructures on land-use conversion decisions within municipal plans. Journal of Transport and Land Use, 7(1), 79. doi:10.5198/jtlu.v7i1.373

Padeiro, M. (2016). Conformance in land-use planning: The determinants of decision, conversion and transgression. Land Use Policy, 55, 285-299. doi:10.1016/j.landusepol.2016.04.014

Parkin, A., & Pugh, C. (1981). Urban Policy and Metropolitan Adeliade. In A. Parkin & A. Patience (Eds.), The Dunstand Decade: Social Democracy at the State Level (pp. 91-114). Melbourne, Australia: Longman Cheshire.

Peduzzi, P., Concato, J., Kemper, E., Holford, T. R., & Feinstein, A. R. (1996). A simulation study of the number of events per variable in logistic regression analysis. Journal of Clinical Epidemiology, 49(12), 1373-1379. doi:10.1016/s0895-4356(96)00236-3

Pendall, R., Wegmann, J., Martin, J., & Wei, D. (2018). The Growth of Control? Changes in Local Land- Use Regulation in Major U.S. Metropolitan Areas From 1994 to 2003. Housing Policy Debate, 1-19. doi:10.1080/10511482.2018.1494024

Phan, T., Peterson, J., & Chandra, S. (2009). Residential Intensificatiopn in a Suburban Fringe Local Government Area, Casey, Melbourne Metropolitan Area, Australia. Australasian Journal of Regional Studies, 15(1), 81-100. Retrieved from: https://search.informit.com.au/fullText;dn=020633366986982;res=IELNZC

PitneyBowes. (2013). MapInfo Business Points Australia: Product Guide. Retrieved from http://www.pbinsight.com.au/support/product-documentation/details/business-points

Planning Workshop Australia. (1996). Major District Centre Scoping Study: draft executive summary. Retrieved from

271

Plantinga, A. J., & Bernell, S. (2007). The Association between Urban Sprawl and Obesity: Is It a Two- Way Street? Journal of Regional Science, 47(5), 857-879. doi:10.1111/j.1467- 9787.2007.00533.x

Pritchard, D. (2002). Are Economic Decisions Rational? Path Dependence, Lock-In and ‘Hinge’ Propositions. Philosophy of Management, 2(3), 29-40. doi:10.5840/pom20022313

Property Council of Australia. (2016). Shopping Centres Online. Retrieved from www.ShoppingCentresOnline.com.au

Prud'homme, R., & Lee, C.-W. (1999). Size, Sprawl, Speed and the Efficiency of Cities. Urban Studies, 36(11), 1849-1858. doi:10.1080/0042098992638

Quigley, J., & Rosenthal, L. (2005). The Effects of Land Use Regulation on the Price of Housing: What Do We Know? What Can We Learn? Cityscape, 8(1), 69-137. Retrieved from: https://www.jstor.org/stable/20868572?seq=1#page_scan_tab_contents

Quirk, M. (2007). Effective centres—a planning dream? Australian Planner, 44(3), 22-29. doi:10.1080/07293682.2007.9982585

Regional Planning Advisory Group, (1991). Princples, Issues & Objectives: A Compilation of Submissions Received in Relation to the SEQ 2001 Conference December 1990.

Reibel, M. (2007). Geographic Information Systems and Spatial Data Processing in Demography: a Review. Population Research and Policy Review, 26(5-6), 601-618. doi:10.1007/s11113-007- 9046-5

Reibel, M., & Agrawal, A. (2007). Areal Interpolation of Population Counts Using Pre-classified Land Cover Data. Population Research and Policy Review, 26(5-6), 619-633. doi:10.1007/s11113- 007-9050-9

REIQ. (1996). Annual 1996 Research Review. Real Estate Institute of Queensland, Brisbane. Retrieved from QLD State Library

Reynolds, H. (1974). Nonparametric Partial Correlation and Causal Analysis. Sociological Methods and Research, 2(3), 376-392. doi:10.1177/004912417400200306

Robertson, J. (2018). Ipswich Council loses $80 million through secretive development arm, ABC News. Retrieved from https://www.abc.net.au/news/2018-06-28/ipswich-council-loses-$80m- through-secretive-development-arm/9914484

272

Rowley, S., & Phibbs, P. (2012). Delivering Diverse and Affordable Housing on Infill Development Sites. Australian Housing and Urban Research Institute, Melbourne. Retrieved from https://www.ahuri.edu.au/research/final-reports/193

Ruming, K. J. (2014). Urban consolidation, strategic planning and community opposition in Sydney, Australia: Unpacking policy knowledge and public perceptions. Land Use Policy, 39, 254-265. doi:10.1016/j.landusepol.2014.02.010

Rundle, A. G., Bader, M. D., Richards, C. A., Neckerman, K. M., & Teitler, J. O. (2011). Using Google Street View to audit neighborhood environments. American Journal of Preventive Medicine, 40(1), 94-100. doi:10.1016/j.amepre.2010.09.034

Sager, T. (2009). Planners' Role: Torn between Dialogical Ideals and Neo-liberal Realities. European Planning Studies, 17(1), 65-84. doi:10.1080/09654310802513948

Sandercock, L. (2005). Picking the Paradoxes: A Historical Anatomy of Australian Planning Cultures. In B. Sanyal (Ed.), Comparative Planning Cultures (pp. 309-330). New York, NY: Routledge.

Schroeder, J. P. (2007). Target-Density Weighting Interpolation and Uncertainty Evaluation for Temporal Analysis of Census Data. Geographical Analysis, 39(3), 311-335. doi:10.1111/j.1538- 4632.2007.00706.x

Schroeder, J. P. (2017). Hybrid Areal Interpolation of Census Counts from 2000 Blocks to 2010 Geographies. Computers, Environment and Urban Systems, 62, 53-63. doi:10.1016/j.compenvurbsys.2016.10.001

Sciara, G. (2015). Measuring Land Uses Performance: Policy, Plan, and Outcome. National Center for Sustainable Transportation, Retrieved from https://escholarship.org/content/qt9w64r1qz/qt9w64r1qz.pdf

Searle, G. (2004). The limits to urban consolidation. Australian Planner, 41(1), 42-48. doi:10.1080/07293682.2004.9982332

Searle, G. (2007). Sydney’s urban consolidation experience: Power, politics and community. Griffith University: Urban Research Program, Brisbane, Australia. Retrieved from http://www.francescoandreone.com/uploads/4/0/0/6/4006916/paper_-_searle_glen_- _sydneys_urban_consolidation_experience-_power_politics_and_community_- _march_2007.pdf

Searle, G. (2010). Too concentrated? The planned distribution of residential density in SEQ. Australian Planner, 47(3), 135-141. doi:10.1080/07293682.2010.508201

273

Seasons, M. (2003). Monitoring and Evaluation in Municipal Planning:Considering the Realities. Journal of the American Planning Association, 69(4), 430-440. doi:10.1080/01944360308976329

SEQROC, Department of Local Government and Planning. (2001). Implementing the South East Queensland Regional Framegwork for Gorwth Management 2000 through local government planning schemes: RFGM guideline no. 1.

Sharam, A., Bryant, L., & Alves, T. (2015). De-risking development of medium density housing to improve housing affordability and boost supply. Australian Planner, 52(3), 210-218. doi:10.1080/07293682.2015.1034146

Siegel, S., & Castellan, N. (1988). Nonparametric Statistics for the Behaviorial Sciences.

Simons, P. L., & Lonergan, N. G. (1973). The Mythical Arguments for Decentralisation. Royal Australian Planning Institute Journal, 11(3), 85-90. doi:10.1080/00049999.1973.9656400

SNAMUTS. (2016). Spatial Network Analysis for Multimodal Urban Transport Systems. http://www.snamuts.com/

Song, Y., & Knaap, G.-J. (2004). Measuring Urban Form: Is Portland Winning the War on Sprawl? Journal of the American Planning Association, 70(2), 210-225. doi:10.1080/01944360408976371

Song, Y., Merlin, L., & Rodriguez, D. (2013). Comparing measures of urban land use mix. Computers, Environment and Urban Systems, 42, 1-13. doi:10.1016/j.compenvurbsys.2013.08.001

Song, Y., & Quercia, R. G. (2008). How are neighbourhood design features valued across different neighbourhood types? Journal of Housing and the Built Environment, 23(4), 297-316. doi:10.1007/s10901-008-9122-0

Stathakis, D., & Tsilimigkas, G. (2014). Measuring the compactness of European medium-sized cities by spatial metrics based on fused data sets. International Journal of Image and Data Fusion, 6(1), 42-64. doi:10.1080/19479832.2014.941018

Steele, W. (2009). Australian Urban Planners: Hybrid Roles and Professional Dilemmas? Urban Policy and Research, 27(2), 189-203. doi:10.1080/08111140902908873

Talen, E. (1996). After the Plans: Methods to Evaluate the Implementation Success of Plans. Journal of Planning Education and Research, 16(2), 79-91. doi:10.1177/0739456x9601600201

274

Talen, E. (1997). Success, failure, and conformance: an alternative approach to planning evaluation. Environment and Planning B: Planning and Design, 24(4), 573-587. doi:10.1068/b240573

Talen, E., Anselin, L., Lee, S., & Koschinsky, J. (2016). Looking for logic: The zoning—land use mismatch. Landscape and Urban Planning, 152, 27-38. doi:10.1016/j.landurbplan.2016.04.002

Talen, E., & Jeong, H. (2018). Does the classic American main street still exist? An exploratory look. Journal of Urban Design, 1-21. doi:10.1080/13574809.2018.1436962

Tanaguchi, M., Matsunaka, R., & Nakamichi, K. (2008). A time-series analysis of relationship between urban layout and automobile reliance: have cities shifted to integration of land use and transport? In C. A. Brebbia (Ed.), Urban Transport XIV: Urban Transport and the Environment in the 21st Century (pp. 415-424). Southampton, UK: WIT Press.

Taşan-Kok, T. (2012). Contradictions of Neoliberal Urban Planning. In T. Taşan-Kok & G. Baeten (Eds.), Contradictions of Neoliberal Planning: Cities, Policies, and Politics (pp. 1-20). Netherlands: Springer.

Terrill, M., Hugh, B., Ha, J., & Hourani, D. (2018). Remarkably adaptive Australian cities in a time of growth. Grattan Institute, Retrieved from https://grattan.edu.au/report/remarkably- adaptive/

The State of Queensland, Regional PLanning Advisory Group. (1993a). Creating our Future: Towards a framework for growth management in South East Queensland.

The State of Queensland, Regional Planning Advisory Group. (1993b). Major Centres: A policy paper of the SEQ2001 Project.

The State of Queensland, Regional Planning Advisory Group. (1993c). The Preferred Pattern or Urban Development for South East Queensland.

The State of Queensland, Regional PLanning Advisory Group. (1993d). Residential Choice and Efficiency: A policy paper of the SEQ2001 Project.

The State of Queensland, Regional Planning Advisory Group. (1994a). Implementing Major Centres Policy.

The State of Queensland, Department of Housing Local Government and Planning. (1994b). Managing Growth in South East Queensland: A Summary of SEQ 2001 recommendations from the Regional Framework for Growth Management.

The State of Queensland, Regional Coordination Committee SEQ 2001 Project. (1995). South East Queensland Regional Framework for Growth Management 1995.

275

The State of Queensland, Regional Coordination Committee SEQ 2001 Project. (1996). South East Queensland Regional Framework for Growth Management Update 1996.

The State of Queensland, Regional Coordination Committee SEQ 2001 Project. (1998). South East Queensland Regional Framework for Growth Management 1998.

The State of Queensland, Regional Coordination Committee SEQ 2001 Project. (2000). South East Queensland Regional Framework for Growth Management 2000.

The State of Queensland, Office of Urban Management. (2004). Draft South East Queensland Regional Plan: For Consultation.

The State of Queensland, Office of Urban Management. (2005a). Consultation Report on the Draft South East Queensland Regional Plan.

The State of Queensland, Office of Urban Management. (2005b). South East Queensland Infrastructure Plan and Program 2005-2026.

The State of Queensland, Office of Urban Management. (2005c). South East Queensland Regional Plan 2005-2026.

The State of Queensland, Office of Urban Management. (2006a). South East Queensland Regional Plan 2005-2026 Amendment 1.

The State of Queensland, Department of Local Government Planning Sport and Recreation. (2006b). Urban Development Monitoring Report: South East Queensland Regional Plan 2005-2026.

The State of Queensland, Department of Infrastructure and Planning. (2008). South East Queensland State of the Region: Technical Report.

The State of Queensland, Queensland Department of Infrastructure and Planning. (2009). South East Queensland Regional Plan 2009-2031.

The State of Queensland, Department of Local Government and Planning. (2010). South East Queensland Growth Management Program Annual Report.

The State of Queensland, Department of Infrastructure Local Government and Planning. (2016a). ShapingSEQ: A summary of community conversations.

The State of Queensland, Department of Infrastructure Local Government and Planning. (2016b). ShapingSEQ: Draft South East Queensland Regional Plan.

276

The State of Queensland, Department of Infrastructure Local Government and Planning. (2017a). ShapingSEQ Background Paper 1: Grow Theme.

The State of Queensland, Department of Infrastructure Local Government and Planning. (2017b). ShapingSEQ Background Paper 2: Prosper Theme.

The State of Queensland, Department of Infrastructure Local Government and Planning. (2017c). ShapingSEQ: South East Queensland Regional Plan 2017.

The State of Queensland, Department of Infrastructure Local Government and Planning. (2017d). South East Queensland Regional Plan 2017: ShapingSEQ Consultation Report.

The State of Queensland. (2018). QImagery. https://qimagery.information.qld.gov.au/

The State of Queensland, & Ipswich City Council, (2007). Ispwich Regional Centre Strategy.

Troy, P. (1996). The Perils of Urban Consolidation. Annandale, NSW: The Federation Press.

Urbis, Department of Infrastructure and Planning and the Ipswich City Council. (2008). Ipswich Regional Centre Strategy Masterplan.

Vanwolleghem, G., Van Dyck, D., Ducheyne, F., De Bourdeaudhuij, I., & Cardon, G. (2014). Assessing the environmental characteristics of cycling routes to school: a study on the reliability and validity of a Google Street View-based audit. International Journal of Health Geographics, 13, 19. doi:10.1186/1476-072X-13-19

von Eye, A., & Mun, E. (2005). Analyzing Rater Agreement: Manifest Variable Methods. New Jersey: Lawerence Erlbaum Associates.

Wildavsky, A. (1973). If planning is everything, maybe it's nothing. Policy Sciences, 4(2), 127-153. doi:10.1007/bf01405729

Williams, K. (1999). Urban intensification policies in England: problems and contradictions. Land Use Policy, 16(3), 167-178. doi:10.1016/s0264-8377(99)00010-1

Willingham, R. (2018). Victoria gets $2.86 billion in 'outstanding deal' to privatise Land Titles and Registry office, ABC. Retrieved from https://www.abc.net.au/news/2018-08-27/victoria- privatises-its-land-titles-and-registry-office/10169056

277

Woodcock, I., Dovey, K., Wollan, S., & Robertson, I. (2011). Speculation and Resistance: Constraints on Compact City Policy Implementation in Melbourne. Urban Policy and Research, 29(4), 343- 362. doi:10.1080/08111146.2011.581335

Wu, L., & Brynjolfsson, E. (2015). The Future of Prediction: How Google Searches Foreshadow Housing Prices and Sales. In A. Goldfarb, S. Greenstein & C. Tucker (Eds.), Economic Analysis of the Digital Economy (pp. 89-118). Chicago: University of Chicago Press.

Yamashita, H., Fujii, T., & Itoh, S. (2006). The Development of Diverse Suburban Activity Cenres in Melbourne, Australia. Applied GIS, 2(2). Retrieved from: http://www.publishing.monash.edu/epress/ag/ag060009.pdf

Yang, Y. (2008). A tale of two cities: Physical form and neighborhood satisfaction in Metropolital Portland and Charlotte. Journal of the American Planning Association, 74(3), 307-323. doi:10.1080/01944360802215546

Zhong, T., Mitchell, B., & Huang, X. (2014). Success or failure: Evaluating the implementation of China's National General Land Use Plan (1997–2010). Habitat International, 44, 93-101. doi:10.1016/j.habitatint.2014.05.003

278

10. Appendices

10.1. Appendix 1 – Input data and manual adjustment of walkable catchment data

This appendix details the rationale for selecting centre transit nodes for each centre. These nodes were used to generate walkable catchment extents using the neighbourhood generator tool by AURIN (2016) as described in section 4.2.1 (page 76). Transit stops were sourced from the South East Queensland general transit feed specifications (https://data.qld.gov.au/dataset/general-transit-feed-specification-gtfs- seq/resource/be7f19e5-3ee8-4396-b9eb-46f6b4ce8039). The rationale for the selection of central transit nodes is detailed in Table 52.

This appendix also describes the manual adjustments necessary to the outputs of the neighbourhood generator, in order to correctly account for possible pedestrian movements.

Table 52 - Rationale for locating central transit nodes in each centre

Centre Rationale for public transit node Name Beenleigh has a bus station and train station with express services located to the SE of the main shopping street. The mean centre (using ArcMap Tools) between the train and bus station platforms has been Beenleigh used as a node for the southern side of the railway. An additional node was defined on the other side of the railway line to account for access from the northern side of the station. The Hyperdome shopping centre features an integrated bus station as Logan the primary location for public transit in the area. The various stops in Hyperdome this station were selected and a point of the mean centre (using ArcMap Tools) was created as the central node. Public transport in Springwood is focussed on its central bus station. The various stops in this station were selected and a point of the mean Springwood centre (using ArcMap Tools) was created as the central node. This node was then moved slightly to the east to sit over the identified road network so as not to confuse the neighbourhood generator tool Logan Central is somewhat unusual in that it a has train-station adjacent to shopping street, and a bus station approximately 1.2km away adjacent to a 1km long strip of big box style shopping centres. The SEQRP identifies the activity centre itself sitting over the Logan Logan City Council office (a location with poor public transit). The train station Central and shopping area at Woodridge represent the most consolidated node of activity and the station has therefore been selected as the central node. A corresponding node was defined on the southern side of the railway line to account for pedestrian accessibility to either side of the railway via the stations The SEQRP shows the Browns Plains centre located in an area of mostly low density residential, approx. 1.2km from the shopping centre Browns that represents the primary node of activity. The shopping centre has Plains an integrated bus station. The various stops in this station were selected and a point of the mean centre (using ArcMap Tools) was created as the central node.

279

Centre Rationale for public transit node Name Springfield is a large scale, master planned, greenfield development. The designated centre is somewhat removed from the majority of the low density residential areas. Most current development for the centre is on the southern side of new railway line. Nodes were created on the southern and northern side of the railway line at Springfield Central Railway station. The southern node was created from the mean centre Springfield the train station platforms and integrated bus stop (using ArcMap Tools), and then moved directly across to the southern side of the railway line. The northern point was located slightly to the east of the station entrance to account for a pedestrian crossing over a local creek. Springfield will require some manual adjustment of the walkability catchment as it does not account for the ability of pedestrians to move the central shopping centre. Ipswich city centre has a conjoined bus and train station in the city centre. There is a single entrance on Bell Street to these facilities. The central transit node was therefore placed at the train station entrance Ipswich on Bell Street. Manual adjustment will be required to walkable catchment to account for new pedestrian route across the Bremer River, and also for pedestrian access to the Ipswich Mall. Goodna train station is separated from the main centre of Goodna by the Ipswich motorway. Pedestrians can access the station from the south via an overpass across the Ipswich motorway. The overpass is approximately 150m long from the platform to Brisbane Road. A central Goodna transit node was located on the northern entrance to the station, and an additional node was defined at the Brisbane Road entrance to the pedestrian overpass. A separate walkability catchment is to be created from this southern node featuring a 650m catchment with 1050m buffer area to account for the length of pedestrian overpass. This centre is based around Garden City Shopping Centre which includes an integrated bus station connected to the SE Busway. The Upper various stops in this station were selected and a point of the mean Mount centre (using ArcMap Tools) was created as the central node. This Gravatt node was then moved slightly to the south to sit over the identified road network so as not to confuse the neighbourhood generator tool Cleveland has a train station with a single main entrance on its eastern Cleveland end. A single central transit node was defined at this entrance Capalaba has a central bus station made up of a plaza surrounded a number of bus stops. The various stops in this station were selected Capalaba and a point of the mean centre (using ArcMap Tools) was created as the central node. This centre is based around Carindale Shopping Centre which includes an integrated bus station. The various stops in this station were Carindale selected and a point of the mean centre (using ArcMap Tools) was created as the central node. Moved the node slightly to the east to sit on the established street network Indooroopilly has both a bus station integrated with a large scale big box shopping centre, and a train station, separated by approximately 400m. Both offer frequent public transport options. The mean centre Indooroopilly (using ArcMap Tools) of the stops at both these facilities was defined as the central transit node. Some manual adjustment of the walkability catchment will be required to account for the pedestrian underpass that permits access to the eastern side of the railway line. Toowong Village Shopping Centre has a train station. Access to this station is via Sherwood Rd or Lissner Street. A node on each of these Toowong streets to account for the virtual barrier created by the shopping centre itself.

280

Centre Rationale for public transit node Name The primary public transport service is the Wynnum Central train Wynnum station. Access is from a ramp on Ronald Street or from an overpass Central on Tingal Road. A central transit node has been placed at each location. Mitchelton includes a train station and a bus station that is integrated with Brookside shopping centre. The mean centre (using ArcMap Tools) of the stops at both these facilities was defined as the central Mitchelton transit node. Some manual adjustment of the walkability catchment will be required to account for the pedestrian access to the southern side of the railway line via dedicated pedestrian routes and the train station. Centro Toombul Shopping Centre has an integrated bus station in proximity to Toombul train station. The mean centre (using ArcMap Tools) of the stops at both these facilities was defined as the central Toombul transit node. An additional node was defined on the western side of the railway to account for pedestrian access across the railway station’s overpass. The Chermside shopping centre features an integrated bus station as the primary location for public transit in the area. The various stops in Chermside this station were selected and a point of the mean center (using ArcMap Tools) was created as the central node. The main point of activity is the Westfield Strathpine Shopping Centre which is approximately 600m from Strathpine train station. The mean centre (using ArcMap Tools) of the stops at both these facilities was Strathpine defined as the central transit node. Some manual adjustment of the walkability catchment will be required to account for pedestrian access to the western side of the railway line via dedicated pedestrian routes and overpasses. Redcliffe is a difficult centre to assess for public transport centrality. Most of the key concentration of centre uses is along a 3km strip on Anzac Ave. High density residential tends to focus on the amenity provided by the bayside and therefore is not necessarily related to the provision of other services. The new railway line (opened 2016) is at the western end of Anzac avenue and will only have had recent impact Redcliffe on development on the area (the planned intent for this station however may have had an impact on determining the location of land uses). The SEQRP designates the centre location approximately mid-way along Anzac Ave. The mean centre (using ArcMap Tools) of all public transport stops on the peninsular, as well as along Anzac Ave, creates a centre approximately half-way along Anzac Ave and has been used as the central node for the centre. Public transit in North Lakes is primarily serviced by its bus station and Mango Hill train station. The train station however did not open until 2016 and its services therefore did not have impact on the overall development of the area (the planned intent for this station however may have had an impact on determining the location of land uses). It appears that most of the land in proximity to the station is already North Lakes developed for low density residential uses. Higher density uses are more proximate to the North Lakes shopping centre and are within the walkable catchment of the bus station. The mean centre (using ArcMap Tools) of the bus station stops has therefore been selected as the central transit node as this reflects the actual delivered public transport services over the past 20 years, and is more related to the designated centre location in SEQRP.

The following manual adjustments were made to the output catchment areas for the centres described below:

281

• Springwood – extended 800m and 1200m catchments using Arcmap ruler and freehand polygon tool to include pedestrian access via pedestrian walkways that were not included in the PSMA Street Network dataset. Also trimmed catchments to avoid pedestrian access to motorway.

• Ipswich - extended 800m and 1200m catchments using Arcmap ruler and freehand polygon tool to include pedestrian access via pedestrian walkways that were not included in the PSMA Street Network dataset (primarily accounting for pedestrian access across the Bremer River).

• Cleveland - extended 800m and 1200m catchments using Arcmap ruler and freehand polygon tool to include pedestrian access via pedestrian walkways that were not included in the PSMA Street Network.

• Toowong – trimmed some of the catchment that extended along ferry lines in the Brisbane River.

• Mitchelton - extended 800m and 1200m catchments using Arcmap ruler and freehand polygon tool to include pedestrian access via pedestrian walkways that were not included in the PSMA Street Network.

• Strathpine - extended 800m and 1200m catchments using Arcmap ruler and freehand polygon tool to include pedestrian access via pedestrian walkways that were not included in the PSMA Street Network.

• Browns Plains - extended 800m and 1200m catchments using Arcmap ruler and freehand polygon tool to include pedestrian access via pedestrian walkways that were not included in the PSMA Street Network.

• Beenleigh – extended 1200m catchments using Arcmap ruler and freehand polygon tool to include pedestrian access via pedestrian walkways that were not included in the PSMA Street Network.

282

10.2. Appendix 2 – Compactness indicator data tables

This appendix provides data tables that describe the results of the various measures used to calculate centre compactness in Chapter 5. The tables are arranged by indicator category.

10.2.1. Residential density Table 53 - 1996 and 2016 Population and Dwelling Densities expressed in net residential hectares

Location Centre Name 1996 2016 1996 2016 Net Pop Net Pop Net Net Density Density Dwelling Dwelling Density Density Inner Toowong 56.9 82.7 31.6 43.2 Carindale 39.1 40.9 14.4 17.1 Chermside 41.0 67.2 22.2 36.8 Indooroopilly 41.3 62.8 19.6 28.6 Middle Mitchelton 34.2 42.7 15.0 19.7 Toombul 50.3 76.8 28.4 42.3 Upper Mount Gravatt 39.3 48.8 15.6 19.2 Wynnum Central 37.2 45.5 18.1 21.5 Beenleigh 33.5 39.3 18.6 20.2 Browns Plains 32.6 32.0 10.8 11.5 Capalaba 36.3 47.1 17.3 25.3 Cleveland 32.6 43.3 15.4 23.5 Goodna 24.3 25.1 9.3 9.4 Ipswich 30.6 32.5 15.3 17.2 Outer Logan Central 40.1 45.9 18.5 19.0 Logan Hyperdome 23.6 23.9 8.4 9.1 North Lakes42 N/A 56.5 N/A 27.3 Redcliffe 38.8 40.4 17.0 19.2 Springfield43 N/A 76.9 N/A 27.6 Springwood 43.3 42.0 17.3 17.0 Strathpine 44.9 46.6 21.4 23.2 44

Table 54 - Estimated centre population and dwelling numbers in 2016 and 1996

Location Centre Name 1996 2016 1996 2016 Population Population Dwellings Dwellings Inner Toowong 7033 10375 3905 5422 Middle Carindale 3782 4186 1395 1754

42 North Lakes and Springfield are greenfield centres and were undeveloped in 1996. 43 44 Land area was determined using the land use database. For low density dwellings (detached houses and duplexes), dwelling numbers were determined by counting the number of building footprints for these uses (duplex counts were doubled to reflect two dwellings occupying a single footprint). Population for low density uses was determined by multiplying the dwelling footprint counts (minus the estimated number of unoccupied dwellings) by the average number of persons per low density dwellings for each census geographic area for the centre (2016 SA1, 1996 CD). The population and dwelling numbers for other residential uses are estimates derived from dasymetric areal interpolation of census data as described in section 4.2.

283

Location Centre Name 1996 2016 1996 2016 Population Population Dwellings Dwellings Chermside 4333 7673 2345 4198 Indooroopilly 5144 7684 2439 3497 Mitchelton 4071 5286 1790 2435 Toombul 6083 9300 3432 5117 Upper Mount Gravatt 3883 4759 1538 1874 Wynnum Central 5759 7073 2802 3337 Beenleigh 2712 3035 1504 1558 Browns Plains 2353 2297 781 825 Capalaba 1390 1592 663 857 Cleveland 2145 3383 1011 1832 Goodna 2660 2738 1019 1032 Ipswich 2035 1904 1016 1006 Outer Logan Central 5780 6594 2665 2726 Logan Hyperdome 3482 3320 1235 1258 North Lakes N/A 2912 N/A 1407 Redcliffe 5351 5513 2353 2620 Springfield N/A 1290 N/A 463 Springwood 2874 2714 1148 1101 Strathpine 1865 2448 889 1221

Table 55 - Relative change to baseline population, dwellings, and densities 1996 to 2016

Relative Relative Relative Relative baseline baseline baseline baseline population dwelling dwelling Baseline population density change density location change change change Inner 46% 46.9% 41% 41.8% Middle 23% 17.0% 23% 17.4% Outer 29% 13.2% 34% 17.9% All 28% 18.3% 30% 20.0%

Table 56 - Relative change in population and dwelling densities 1996 -2016 in terms of built-up hectares45

Location Centre Name Relative change Relative change Dwelling Population Density Density (Dwellings / Built-up (Pop / Built-up Hectare Hectare) Inner Toowong 47.5% 38.8% Carindale 5.9% 20.3% Chermside 75.7% 77.6% Middle Indooroopilly 50.3% 44.3% Mitchelton 29.1% 35.3%

45 These density change figures are calculated in terms of built-up hectares in order to be comparable to baseline density figures. They differ to the net density figures shown for net residential density (see discussion on baselines in section 4.2.2 for further explanation).

284

Location Centre Name Relative change Relative change Dwelling Population Density Density (Dwellings / Built-up (Pop / Built-up Hectare Hectare) Toombul 48.5% 44.9% Upper Mount Gravatt 19.2% 18.6% Wynnum Central 22.8% 19.1% Beenleigh 11.9% 3.6% Browns Plains -12.8% -5.6% Capalaba 14.5% 29.3% Cleveland 57.7% 81.2% Goodna -0.7% -2.3% Outer Ipswich -6.5% -1.0% Logan Central 14.1% 2.3% Logan Hyperdome -4.7% 1.8% Redcliffe 3.0% 11.3% Springwood -5.6% -4.0% Strathpine 16.3% 21.7%

Table 57 - Residential density indicators for 2016 and 1996 expressed as z-Scores (ordered by 2016 score)

Score

Density

low densities low

Proportion of Proportion

Net Population Population Net

Average Density Density Average

population living at at living population

Net Dwelling Density Dwelling Net Average Land area of area Land Average dwellings density low Location Centre 2016 1996 2016 1996 2016 1996 2016 1996 2016 1996 Inner Toowong 1.62 1.91 2.08 2.37 2.20 2.46 0.52 0.59 1.68 2.20 Middle Toombul 1.49 1.41 1.72 1.55 2.10 1.89 0.72 0.85 1.43 1.35 Middle Chermside 1.15 0.60 1.14 0.39 1.51 0.81 0.62 0.90 1.35 0.30 Outer Springfield 0.72 N/A 1.73 N/A 0.52 N/A 1.83 N/A -1.21 N/A Outer North Lakes 0.60 N/A 0.49 N/A 0.49 N/A 0.81 N/A 0.59 N/A Middle Indooroopilly 0.58 0.39 0.87 0.43 0.63 0.35 -0.17 -0.13 1.00 0.92 Outer Strathpine 0.36 0.83 -0.12 0.87 0.05 0.67 0.35 0.69 1.17 1.09 Outer Capalaba 0.29 -0.03 -0.09 -0.20 0.28 -0.05 -0.15 -1.22 1.13 1.36 Outer Cleveland 0.04 -0.31 -0.32 -0.66 0.08 -0.39 -0.20 -0.17 0.60 -0.02 Middle Upper Mount Gravatt -0.09 -0.09 0.02 0.18 -0.38 -0.35 0.34 0.63 -0.33 -0.82 Middle Wynnum Central -0.09 0.02 -0.18 -0.09 -0.14 0.09 0.71 0.70 -0.74 -0.63 Outer Logan Central -0.18 0.20 -0.16 0.27 -0.41 0.16 -0.14 0.17 0.00 0.21 Middle Mitchelton -0.24 -0.34 -0.35 -0.46 -0.33 -0.45 0.51 0.61 -0.80 -1.08 Outer Beenleigh -0.25 -0.07 -0.56 -0.55 -0.27 0.17 -0.52 -0.34 0.33 0.44 Outer Redcliffe -0.33 -0.02 -0.49 0.11 -0.38 -0.10 0.52 0.83 -0.97 -0.93 Outer Springwood -0.37 0.22 -0.40 0.67 -0.62 -0.05 0.06 0.44 -0.51 -0.16 Middle Carindale -0.38 -0.19 -0.46 0.15 -0.61 -0.56 0.13 0.52 -0.56 -0.88 Outer Ipswich -0.63 -0.46 -0.97 -0.90 -0.60 -0.40 -0.20 0.06 -0.75 -0.60 Outer Browns Plains -1.10 -0.93 -1.01 -0.66 -1.21 -1.19 -0.80 -0.49 -1.39 -1.38

285

Outer Goodna -1.50 -1.40 -1.43 -1.69 -1.43 -1.45 -2.10 -1.80 -1.05 -0.67 Outer Logan Hyperdome -1.70 -1.74 -1.50 -1.78 -1.48 -1.62 -2.84 -2.84 -0.97 -0.71

Table 58 - Difference in residential density indicators, and relative estimated population and dwelling change

1996 to 2016, expressed as z-Scores (ordered by score)

Dwelling Dwelling

density density intensification z score Relative population change Relative dwelling change Net Population Density Change Change Net Density Change land Average of low area density dwellings change of Proportion population low living at densities change Location Centre Average Middle Chermside 1.74 2.23 2.11 1.85 2.09 -0.52 2.67 Middle Toombul 1.13 1.24 0.93 1.88 1.94 -0.02 0.84 Outer Cleveland 1.05 1.44 2.19 0.21 0.72 0.34 1.42 Middle Indooroopilly 0.83 1.10 0.71 1.34 0.91 0.32 0.63 Inner Toowong 0.79 1.02 0.53 1.80 1.46 0.14 -0.19 Outer Capalaba 0.76 -0.33 0.15 0.22 0.71 3.74 0.08 Middle Mitchelton 0.13 0.30 0.42 -0.02 -0.01 0.07 0.02 Upper Mount Middle -0.01 0.00 -0.14 0.08 -0.22 -0.50 0.71 Gravatt Outer Strathpine -0.08 0.36 0.47 -0.74 -0.60 -0.67 0.69 Middle Wynnum Central -0.14 0.01 -0.25 -0.04 -0.27 0.37 -0.69 Middle Carindale -0.36 -0.49 0.01 -0.73 -0.41 -0.78 0.23 Outer Beenleigh -0.40 -0.44 -0.85 -0.30 -0.64 -0.08 -0.12 Outer Logan Central -0.58 -0.35 -0.91 -0.31 -0.89 -0.55 -0.49 Outer Redcliffe -0.64 -0.80 -0.55 -0.74 -0.53 -0.57 -0.67 Outer Logan Hyperdome -0.71 -1.11 -0.92 -0.89 -0.84 0.63 -1.10 Outer Ipswich -0.78 -1.19 -1.03 -0.72 -0.59 -0.37 -0.76 Outer Browns Plains -0.82 -1.02 -0.77 -0.98 -0.84 -0.45 -0.85 Outer Goodna -0.88 -0.80 -0.95 -0.84 -0.96 -0.33 -1.38 Outer Springwood -1.04 -1.15 -1.15 -1.06 -1.04 -0.77 -1.04

286

10.2.2. Dwelling mix Figure 55 - Proportions of dwelling types, 2016 and 1996

287

Table 59 - Change in proportions of dwelling types, 1996 to 2016

Change in Change in Change in Change in Location Centre LD Prop LMD Prop MD Prop HD Prop Inner Toowong -0.10 -0.07 0.05 0.13 Carindale -0.17 0.11 0.00 0.05 Chermside -0.34 0.12 -0.05 0.28 Indooroopilly -0.16 -0.03 0.03 0.16 Middle Mitchelton -0.18 0.07 0.09 0.03 Toombul -0.17 -0.04 0.04 0.16 Upper Mount Gravatt -0.18 0.03 0.03 0.13 Wynnum Central -0.06 0.04 0.02 0.00 Beenleigh -0.05 -0.03 0.08 0.00 Browns Plains -0.05 0.05 0.00 0.00 Capalaba -0.11 0.10 0.01 0.00 Cleveland -0.26 0.14 0.06 0.06 Goodna 0.01 0.00 -0.01 0.00 Outer Ipswich -0.08 0.02 0.03 0.02 Logan Central -0.02 0.01 0.00 0.01 Logan Hyperdome 0.00 0.00 0.00 0.00 Redcliffe -0.09 -0.02 0.03 0.08 Springwood 0.01 -0.01 0.00 0.00 Strathpine -0.11 0.10 0.01 0.00 Inner -0.18 -0.05 0.05 0.19 Middle -0.09 0.06 0.01 0.01 Baseline Outer -0.05 0.04 0.00 0.01 All -0.09 0.04 0.01 0.04

Table 60 - Absolute change in dwelling numbers by type of dwelling, 1996 to 2016

Absolute LD Absolute LMD Absolute MD Absolute HD Dwelling Dwelling Dwelling Dwelling Change Change Change Change Location Centre Inner Toowong -139 -21 696 981 Carindale 12 255 0 92 Chermside -340 928 96 1168 Indooroopilly -55 130 354 629 Middle Mitchelton 124 229 221 70 Toombul -195 349 707 824 Upper Mount Gravatt -63 107 57 235 Wynnum Central 200 232 89 13 Beenleigh -50 -17 121 0 Browns Plains 4 40 0 0 Outer Capalaba -26 209 11 0 Cleveland 43 539 133 107

288

Goodna 16 2 -5 0 Ipswich -86 20 33 23 Logan Central -10 44 0 27 Logan Hyperdome 15 7 0 0 Redcliffe -15 -8 73 217 Springwood -19 -27 0 0 Strathpine -5 321 16 0

Inner -2793 1383 6921 18635 Middle 12721 16379 3368 2843 Baseline Outer 32981 13811 629 1416 All 42126 31063 10439 22329

Table 61 - Dwelling mix indicators for 2016 and 1996 expressed as z-Scores (ordered by 2016 score)

score

of LD of LD

score score score of HD score

of of MD

of LMD of LMD

Average Average

dwelling dwelling

Proportion Proportion Proportion Proportion Proportion

Dwellings z Dwellings z Dwellings z Dwellings z Dwellings score z IQV mix z mix

Location Centre 2016 1996 2016 1996 2016 1996 2016 1996 2016 1996 2016 1996

Inner Toowong 1.42 1.90 1.54 1.81 -0.75 -0.22 2.43 2.22 2.52 3.82 1.39 1.88

Middle Toombul 1.20 0.98 1.35 1.28 -0.14 0.24 2.43 2.28 0.93 -0.30 1.43 1.40

Middle Chermside 1.07 0.34 1.24 0.28 0.24 -0.29 0.41 1.16 2.12 -0.30 1.36 0.84

Middle Indooroopilly 1.01 0.94 0.92 0.78 -0.69 -0.44 1.72 1.65 1.66 1.34 1.47 1.39

Outer North Lakes 0.60 N/A 0.79 N/A 0.74 N/A 0.38 N/A 0.13 N/A 0.96 N/A

Outer Cleveland 0.47 0.03 0.72 0.07 0.98 0.41 0.05 -0.38 -0.11 -0.30 0.72 0.33

Outer Capalaba 0.33 0.59 1.23 1.41 2.36 2.17 -0.62 -0.57 -0.70 -0.30 -0.60 0.22

Outer Strathpine 0.32 0.54 1.07 1.21 2.11 1.89 -0.56 -0.49 -0.70 -0.30 -0.34 0.38

Outer Beenleigh 0.22 0.34 0.42 0.76 0.94 1.38 0.00 -0.57 -0.70 -0.30 0.44 0.45

Outer Logan Central -0.09 0.21 0.03 0.47 0.80 1.01 -0.74 -0.57 -0.60 -0.30 0.08 0.42 Upper Mount Middle Gravatt -0.11 -0.61 -0.42 -0.88 -0.71 -0.81 -0.26 -0.32 0.57 -0.30 0.26 -0.74

Outer Ipswich -0.35 -0.37 -0.53 -0.51 -0.20 -0.19 -0.43 -0.57 -0.47 -0.30 -0.11 -0.28

Middle Carindale -0.38 -0.72 -0.57 -0.99 -0.22 -0.78 -0.74 -0.57 -0.17 -0.30 -0.17 -0.95

Middle Mitchelton -0.38 -0.89 -0.68 -1.21 -0.77 -1.10 0.16 -0.51 -0.41 -0.30 -0.18 -1.34 Wynnum Middle Central -0.38 -0.31 -0.60 -0.49 -0.34 -0.44 -0.14 -0.14 -0.66 -0.30 -0.17 -0.15

Outer Springwood -0.43 -0.08 -0.52 -0.06 0.12 0.36 -0.74 -0.57 -0.70 -0.30 -0.29 0.16

Outer Redcliffe -0.53 -0.71 -0.89 -0.99 -0.99 -0.83 -0.42 -0.50 0.14 -0.30 -0.51 -0.95 Logan Outer Hyperdome -0.71 -0.42 -0.92 -0.59 -0.40 -0.29 -0.74 -0.57 -0.70 -0.30 -0.78 -0.37

Outer Goodna -0.83 -0.53 -1.09 -0.76 -0.66 -0.57 -0.68 -0.45 -0.70 -0.30 -1.04 -0.57

Outer Springfield46 -1.20 N/A -1.50 N/A -1.16 N/A -0.74 N/A -0.70 N/A -1.87 N/A

46 Springfield is a greenfield development still subject to more intensification. It is included here for reference, but these current figures are based on the housing within the centre extent which is currently made up entirely of low density dwellings and duplexes. The Springfield master plan allows for a wide range of dwelling types in a number of currently undeveloped locations within the centre extent.

289

Outer Browns Plains -1.27 -1.22 -1.58 -1.59 -1.26 -1.53 -0.74 -0.57 -0.70 -0.30 -2.05 -2.11

Table 62 - Relative change of dwelling types (1996-2016) expressed as z-scores (ordered by overall average z-

score)

score score score score

z scores

Dwelling Dwelling Dwelling Dwelling Dwelling

change z change z change z change z change

Average of of Average

Relative LD LD Relative

Relative HD HD Relative

Relative MD MD Relative

dwelling mix dwelling Relative LMD LMD Relative Location Centre intensification Middle Chermside 1.77 2.58 1.69 -0.23 3.02 Middle Toombul 1.02 0.84 -0.11 2.23 1.10 Inner Toowong 0.67 0.43 -0.77 1.82 1.19 Outer Cleveland 0.66 -1.11 2.54 1.12 0.10 Middle Indooroopilly 0.58 0.17 -0.41 1.33 1.24 Outer Capalaba 0.11 0.50 1.20 -0.60 -0.68 Middle Upper Mount Gravatt 0.10 0.53 -0.31 -0.29 0.46 Outer Strathpine 0.02 -0.16 1.49 -0.57 -0.68 Outer Ipswich -0.02 1.39 -0.62 -0.35 -0.51 Outer Beenleigh -0.19 0.38 -0.80 0.36 -0.68 Middle Mitchelton -0.24 -1.64 0.05 1.00 -0.39 Middle Carindale -0.27 -0.45 0.39 -0.84 -0.19 Outer Redcliffe -0.32 -0.15 -0.76 -0.37 0.01 Outer Logan Central -0.57 -0.20 -0.64 -0.84 -0.61 Outer Browns Plains -0.58 -0.38 -0.42 -0.84 -0.68 Outer Springwood -0.59 0.05 -0.88 -0.84 -0.68 Outer Logan Hyperdome -0.68 -0.51 -0.70 -0.84 -0.68 Outer Goodna -0.73 -0.58 -0.72 -0.92 -0.68 Middle Wynnum Central -0.73 -1.68 -0.23 -0.36 -0.65

10.2.3. Employment Table 63 - Net employment density, 1996 and 2016

Net job Net job density 1996 density 2016 (jobs / net (jobs / net employment employment Location Centre hectare) hectare) Difference Middle Carindale 74 167 93.4 Middle Chermside 85 146 61.4 Middle Upper Mount Gravatt 75 122 47.1 Outer Ipswich 80 122 42.8 Inner Toowong 112 154 41.5 Middle Indooroopilly 62 97 35.5 Outer Cleveland 48 71 23.3 Outer Redcliffe 59 82 23.2 Outer Logan Hyperdome 60 83 22.8

290

Outer Strathpine 49 63 13.5 Outer Logan Central 52 64 11.4 Outer Beenleigh 45 55 10.2 Outer Goodna 30 39 9.4 Middle Wynnum Central 61 70 9.0 Middle Mitchelton 48 57 8.7 Outer Capalaba 67 74 7.2 Outer Springwood 62 69 7.0 Middle Toombul 79 84 4.9 Outer Browns Plains 62 60 -1.7 Outer North Lakes N/A 67 N/A Outer Springfield N/A 75 N/A

Table 64 - Most significant estimated employment change by use and centre, 1996 to 2016

Location Centre Use type Employment increase % of employment change Inner Toowong Hospital 1305 73% Education Facility 177 10% Office 162 9% Middle Carindale Shopping Centre - Box 2326 79% Hospital 411 14% Shopping Centre - Strip 78 3% Chermside Shopping Centre - Box 3047 52% Hospital 2285 39% Mixed use complex 227 4% Indooroopilly Shopping Centre - Box 1735 73% Education Facility 433 18% Office 95 4% Mitchelton Shopping Centre - Box 175 42% Shopping Centre - Strip 92 22% Hospital 80 19% Toombul47 Mixed use residential 400 154% Shopping Centre - Box 64 24% Education Facility 52 20% Upper Mount Shopping Centre - Box 2562 74% Gravatt Office 650 19% Health Care 56 2% Wynnum Central Education Facility 90 61% Office 57 39% Retail 51 34% Outer Beenleigh Shopping Centre - Box 632 61% Office 194 19% Education Facility 164 16% Browns Plains Main Street 1152 28% Bulky Goods Retail 854 21% Warehousing 616 15% Capalaba Bulky Goods Retail 415 31% Shopping Centre - Box 397 29% Office 158 12% Cleveland Shopping Centre - Box 488 31% Office 368 23%

47 Toombul saw a significant reduction in some use types (primarily main street retail) through redevelopment, resulting in this measures exceeding 100%

291

Shopping Centre - Strip 367 23% Goodna Education Facility 109 29% Office 87 23% Shopping Centre - Strip 43 11% Ipswich Shopping Centre - Box 1861 41% Hospital 921 20% Mixed use office 874 19% Logan Central Office 400 57% Education Facility 173 24% Shopping Centre - Strip 90 13% Logan Hyperdome Shopping Centre - Box 1038 63% Bulky Goods Retail 245 15% Office 240 15% North Lakes Shopping Centre - Box 1894 33% Hospital 1019 18% Office 868 15% Redcliffe Hospital 805 56% Shopping Centre - Box 242 17% Health Care 93 6% Springfield Shopping Centre - Box 2081 52% Office 1303 32% Education Facility 369 9% Springwood Office 440 65% Short Term 156 23% Accommodation Shopping Centre - Box 82 12% Strathpine Shopping Centre - Strip 347 24% Shopping Centre - Box 266 18% Office 164 11%

Table 65 - Employment plot ratio, 2016 and 1996, (ordered by difference)

Employment Employment Location Centre Difference plot ratio 2016 plot ratio 1996 Middle Carindale 0.54 0.25 0.30 Middle Chermside 0.37 0.21 0.16 Middle Upper Mount Gravatt 0.42 0.27 0.16 Middle Indooroopilly 0.39 0.25 0.14 Outer Ipswich 0.43 0.28 0.14 Inner Toowong 0.48 0.37 0.11 Outer Cleveland 0.30 0.21 0.09 Outer Logan Hyperdome 0.31 0.23 0.09 Outer Browns Plains 0.31 0.24 0.07 Outer Redcliffe 0.24 0.18 0.06 Outer Strathpine 0.26 0.21 0.06 Outer Capalaba 0.33 0.28 0.05 Outer Logan Central 0.24 0.20 0.05 Middle Wynnum Central 0.32 0.28 0.04 Outer Beenleigh 0.21 0.18 0.04 Outer Goodna 0.16 0.12 0.04 Middle Mitchelton 0.24 0.20 0.03 Outer Springwood 0.33 0.30 0.03 Middle Toombul 0.30 0.29 0.02 Outer North Lakes 0.23 N/A N/A Outer Springfield 0.28 N/A N/A

292

Table 66 – Employment indicators for 2016 and 1996 expressed as z-Scores (ordered by 2016 z score)

score

score score score

Net job job Net

Average Average

density z density

intensity z intensity

plot ratio z plot ratio

employment employment

Employment Employment Employment

Location Centre 2016 1996 2016 1996 2016 1996 2016 1996 Middle Carindale 1.87 0.52 2.29 0.55 0.94 0.88 2.38 0.14 Inner Toowong 1.59 2.00 1.91 2.68 1.10 0.96 1.75 2.35 Middle Chermside 1.43 1.02 1.69 1.15 2.11 2.45 0.49 -0.54 Outer Ipswich 0.91 0.72 1.02 0.88 0.59 0.50 1.13 0.79 Middle Upper Mount Gravatt 0.90 0.54 1.01 0.63 0.59 0.50 1.11 0.51 Middle Indooroopilly 0.26 -0.10 0.29 -0.11 -0.29 -0.34 0.77 0.16 Outer Redcliffe 0.15 0.03 -0.13 -0.25 1.40 1.40 -0.82 -1.05 Middle Toombul 0.05 0.71 -0.07 0.87 0.37 0.41 -0.16 0.85 Outer Logan Hyperdome -0.02 -0.09 -0.11 -0.19 0.10 0.16 -0.06 -0.24 Outer Springfield -0.20 N/A -0.34 N/A 0.14 N/A -0.41 N/A Outer North Lakes -0.29 N/A -0.57 N/A 0.67 N/A -0.97 N/A Outer Cleveland -0.41 -0.79 -0.45 -0.88 -0.55 -0.99 -0.23 -0.52 Outer Capalaba -0.42 0.07 -0.37 0.16 -1.02 -0.76 0.13 0.80 Outer Logan Central -0.47 -0.41 -0.65 -0.62 0.03 0.16 -0.81 -0.76 Middle Wynnum Central -0.52 -0.24 -0.48 -0.14 -1.05 -1.25 -0.04 0.68 Outer Springwood -0.61 -0.23 -0.50 -0.07 -1.46 -1.71 0.12 1.08 Outer Strathpine -0.61 -0.68 -0.68 -0.79 -0.56 -0.72 -0.59 -0.53 Outer Beenleigh -0.69 -0.76 -0.90 -1.03 -0.03 -0.12 -1.14 -1.13 Middle Mitchelton -0.73 -0.73 -0.85 -0.85 -0.45 -0.73 -0.90 -0.61 Outer Browns Plains -1.00 -0.08 -0.76 -0.11 -2.19 -0.22 -0.05 0.09 Outer Goodna -1.16 -1.51 -1.36 -1.88 -0.42 -0.58 -1.71 -2.08

Table 67 - Difference in employment indicators, and relative estimated employment change 1996 to 2016,

expressed as z-Scores (ordered by score)

scores

change change change

Relative Relative

plot ratio plot ratio

Average of of Average

employment employment employment

Employment Employment intensification intensification Location Centre density job Net Middle Carindale 2.81 2.50 2.87 3.06 Middle Chermside 1.07 0.67 1.53 1.02 Middle Upper Mount Gravatt 0.87 0.65 0.93 1.01 Outer Ipswich 0.60 0.24 0.76 0.80 Middle Indooroopilly 0.53 0.31 0.45 0.82 Outer Browns Plains 0.35 2.41 -1.11 -0.25 Inner Toowong 0.21 -0.44 0.70 0.38 Outer Cleveland 0.19 0.61 -0.06 0.02 Outer Logan Hyperdome 0.06 0.26 -0.08 0.01

293

Outer Redcliffe -0.16 -0.06 -0.07 -0.36 Outer Strathpine -0.36 -0.13 -0.47 -0.48 Outer Beenleigh -0.49 -0.14 -0.61 -0.74 Outer Goodna -0.50 -0.11 -0.64 -0.74 Outer Logan Central -0.56 -0.53 -0.56 -0.58 Outer Capalaba -0.57 -0.40 -0.74 -0.58 Middle Mitchelton -0.77 -0.79 -0.67 -0.84 Middle Wynnum Central -0.79 -0.99 -0.66 -0.71 Outer Springwood -0.79 -0.81 -0.74 -0.81 Middle Toombul -0.94 -0.97 -0.83 -1.02

10.2.4. Mixed use

Table 68 - Mixed use indicators 2016 and 1996, expressed as z scores and ordered by 2016 score

score score score score

Active Active

Median Median

Average Average

Land use Land use

Euclidean Euclidean

frontage z z frontage

variation z z variation

residential residential

Residential Residential z Proximity

use z score z use

Difference z z Difference

Proximity of Proximity

Proportional Proportional

uses z score z uses Average mixed mixed Average

Location Centre 2016 1996 2016 1996 2016 1996 2016 1996 2016 1996 2016 1996

Outer Ipswich 1.45 1.53 1.73 1.68 1.26 1.16 1.34 1.66 1.61 1.64 1.32 1.51

Inner Toowong 0.75 0.78 -0.52 -0.30 1.09 0.97 0.95 1.12 1.18 1.45 1.03 0.67

Middle Chermside 0.73 0.62 0.62 1.24 0.58 0.63 0.72 0.68 0.99 0.78 0.74 -0.20

Middle Toombul 0.64 0.55 -0.57 -0.24 0.85 0.84 0.39 0.25 1.30 1.13 1.23 0.74

Outer Cleveland 0.57 0.53 0.12 0.25 0.44 0.30 0.41 0.18 0.32 -0.36 1.55 2.26

Middle Indooroopilly 0.50 0.56 -0.32 -0.02 0.28 0.47 1.23 1.39 0.99 0.88 0.31 0.06

Middle Wynnum Central 0.42 0.38 -1.75 -1.57 0.37 0.49 0.76 0.90 0.99 1.06 1.72 1.02

Outer Beenleigh 0.31 0.21 0.85 0.83 0.67 0.57 -0.62 -1.00 -0.40 -0.58 1.05 1.21

Outer Logan Central 0.15 0.23 -0.04 0.25 0.21 0.28 0.74 0.80 -0.12 -0.22 -0.03 0.06

Middle Mitchelton 0.15 0.25 -0.58 -0.05 0.52 0.51 0.66 0.74 0.18 0.08 -0.02 -0.01

Outer Springwood 0.13 0.06 0.79 0.97 0.37 0.24 -0.05 -0.25 0.62 0.50 -1.05 -1.14

Outer North Lakes 0.02 N/A 1.44 N/A -0.19 N/A -0.27 N/A -0.69 N/A -0.19 N/A

Outer Strathpine -0.12 0.03 0.83 1.33 0.12 -0.01 -0.52 -0.87 -0.46 0.12 -0.58 -0.43

Outer Redcliffe -0.20 -0.22 -1.26 -1.22 0.25 0.32 0.42 0.37 0.19 0.02 -0.61 -0.55

Outer Goodna -0.22 -0.38 -1.17 -1.15 0.08 0.04 0.53 -0.01 0.01 -0.30 -0.57 -0.47

Middle Upper Mount Gravatt -0.32 -0.38 0.54 0.86 0.15 0.23 -0.76 -0.98 -0.73 -1.09 -0.79 -0.93

Outer Capalaba -0.52 -1.02 0.53 0.42 -0.54 -1.66 -0.52 -1.39 -1.48 -2.05 -0.59 -0.44

Middle Carindale -0.80 -0.86 -1.15 -0.82 -0.97 -1.07 -0.55 -0.94 -0.49 -0.81 -0.86 -0.68

Outer Browns Plains -0.84 -1.33 -0.36 -1.03 -1.25 -2.09 -0.71 -0.97 -1.07 -1.20 -0.81 -1.38

Outer Logan Hyperdome -1.06 -1.53 -1.17 -1.42 -1.00 -2.22 -1.11 -1.68 -0.60 -1.03 -1.41 -1.28

Outer Springfield -1.74 N/A 1.43 N/A -3.31 N/A -3.05 N/A -2.34 N/A -1.43 N/A

294

Table 69 - Difference in mixed use indicators expressed as z-scores (ordered by score)

proximity proximity

mixed use use mixed intensification scores in Difference Euclidean difference scores in Difference residential proximity scores in Difference prop score in Difference frontages active score Location Centre of Average Outer Capalaba 0.95 2.18 2.62 -0.61 -0.39 Middle Chermside 0.93 -0.60 0.18 2.13 1.99 Middle Toombul 0.66 -0.56 0.13 2.42 0.66 Outer Logan Hyperdome 0.50 2.53 0.09 -0.55 -0.08 Outer Browns Plains 0.41 1.80 -0.92 -0.79 1.56 Outer Goodna 0.34 -0.27 2.47 -0.58 -0.25 Inner Toowong 0.23 -0.38 -0.58 1.47 0.40 Middle Indooroopilly 0.03 -0.83 -0.09 0.68 0.36 Middle Wynnum Central -0.07 -0.70 -0.72 0.07 1.07 Middle Carindale -0.15 0.15 -0.01 -0.38 -0.37 Outer Springwood -0.19 -0.16 -0.21 -0.78 0.38 Outer Redcliffe -0.33 -0.55 -0.25 -0.36 -0.14 Outer Cleveland -0.37 -0.16 0.66 0.58 -2.55 Outer Strathpine -0.38 -0.10 -0.18 -0.85 -0.38 Middle Mitchelton -0.39 -0.46 -0.62 -0.27 -0.22 Middle Upper Mount Gravatt -0.41 -0.55 -1.22 -0.32 0.44 Outer Logan Central -0.48 -0.55 -0.36 -0.58 -0.42 Outer Beenleigh -0.48 -0.31 -0.16 -0.50 -0.95 Outer North Lakes -0.60 -0.72 -0.86 -0.84 0.02 Outer Springfield -0.60 -0.72 -0.86 -0.84 0.02 Outer Ipswich -0.80 -0.48 -0.81 -0.80 -1.11

10.2.5. Summary tables of overall scores

Table 70 – Overall centre compactness scores, 2016 and 1996

core

compactness compactness s Residential Score Density Mix Dwelling Score Employment Score score use Mixed Overall Location Centre 2016 1996 2016 1996 2016 1996 2016 1996 2016 1996 Inner Toowong 1.34 1.65 1.62 1.91 1.42 1.90 1.59 2.00 0.75 0.78 Middle Chermside 1.10 0.65 1.15 0.60 1.07 0.34 1.43 1.02 0.73 0.62 Middle Toombul 0.84 0.91 1.49 1.41 1.20 0.98 0.05 0.71 0.64 0.55 Middle Indooroopilly 0.59 0.45 0.58 0.39 1.01 0.94 0.26 -0.10 0.50 0.56 Outer Ipswich 0.35 0.36 -0.63 -0.46 -0.35 -0.37 0.91 0.72 1.45 1.53 Outer North Lakes 0.23 N/A 0.60 N/A 0.60 N/A -0.29 N/A 0.02 N/A Outer Cleveland 0.17 -0.14 0.04 -0.31 0.47 0.03 -0.41 -0.79 0.57 0.53

295

Middle Upper Mount Gravatt 0.10 -0.14 -0.09 -0.09 -0.11 -0.61 0.90 0.54 -0.32 -0.38 Middle Carindale 0.08 -0.31 -0.38 -0.19 -0.38 -0.72 1.87 0.52 -0.80 -0.86 Outer Strathpine -0.01 0.18 0.36 0.83 0.32 0.54 -0.61 -0.68 -0.12 0.03 Outer Capalaba -0.08 -0.10 0.29 -0.03 0.33 0.59 -0.42 0.07 -0.52 -1.02 Outer Beenleigh -0.10 -0.07 -0.25 -0.07 0.22 0.34 -0.69 -0.76 0.31 0.21 Middle Wynnum Central -0.14 -0.04 -0.09 0.02 -0.38 -0.31 -0.52 -0.24 0.42 0.38 Outer Logan Central -0.15 0.06 -0.18 0.20 -0.09 0.21 -0.47 -0.41 0.15 0.23 Outer Redcliffe -0.23 -0.23 -0.33 -0.02 -0.53 -0.71 0.15 0.03 -0.20 -0.22 Middle Mitchelton -0.30 -0.43 -0.24 -0.34 -0.38 -0.89 -0.73 -0.73 0.15 0.25 Outer Springwood -0.32 -0.01 -0.37 0.22 -0.43 -0.08 -0.61 -0.23 0.13 0.06 Outer Springfield -0.61 N/A 0.72 N/A -1.20 N/A -0.20 N/A -1.74 N/A Outer Logan Hyperdome -0.87 -0.94 -1.70 -1.74 -0.71 -0.42 -0.02 -0.09 -1.06 -1.53 Outer Goodna -0.93 -0.96 -1.50 -1.40 -0.83 -0.53 -1.16 -1.51 -0.22 -0.38 Outer Browns Plains -1.05 -0.89 -1.10 -0.93 -1.27 -1.22 -1.00 -0.08 -0.84 -1.33

296

10.3. Appendix 3 – Adjustments to shopping centre areas

This appendix describes manual adjustments made to the shopping centre areas recorded in shopping centre directories (Building Owners and Managers Association Queensland Division, 1993; Property Council of Australia, 2016)

2016 • The PCA recorded Village Square in Browns Plains as being 37,000m2. However, when calling centre management they advised that they are only 12,500m2. They believe the PCA have erroneously included the new centre to south as part of their floor area. Uses were merged.

• Indooroopilly Central shopping centre had a large discrepancy between the listed GLA and the building footprint area. After calling centre management, they confirmed that they have a lower level that is not visible from the street. This includes a 9,000m2 self-storage facility. As this is not a common shopping centre use, and has minimal employment, subtracted 9000m2 from the floor area for this centre.

1996 • Grand Plaza shopping centre in Browns Plains - was originally developed in 1994 and is therefore not included in the 1993 directory. It underwent a major expansion in 2006. The approval for this expansion (MCUI-4/2005) states that this extension resulted in an additional floor area of 11,165m2. The 1996 floor area was therefore determined by subtracting this from the 2016 floor area listed in the PCA directory, making a total area of 42,169m2

• Capalaba Centre Shopping Centre – similar to Grand Plaza, this centre was established in 1994, after the 1993 directory. It has a complex history involving Supreme Court interventions. It underwent an expansion in 2008 (MC010486). The planning report for this extension states that the centre was expanded by a total 5551m2. The floor area was therefore determined by subtracting this from the 2016 floor area listed in the PCA directory, making a total area of 35,741m2

• The Latest PCA directory does not list this centre’s GFA. I therefore updated the 2016 floor area to equal the 1993 GLA. There have been alterations to the building since 1993, however these appear to be relatively minor and the centre has primarily seen cosmetic changes.

297

• Indooroopilly Centre – no data existed in the 1993 directory (centre opening in 1994). However, the footprint from the 1996 aerial images appears identical to existing images. The 2016 GFA was therefore applied to the site. (10,767m2)

• Ipswich City Square – this centre had a section redeveloped as an office tower. The 1993 GFA and site is therefore larger than the 2016 site.

• Ipswich City Plaza was not listed in the 1993 directory, however the building appears unchanged, and it was therefore given the 2016 GLA value of 4,855m2

• Bluewater Square in Redcliff – the old centre is not listed in the directory and it was significantly expanded in 2008/9. Absent other data, the old foorprint area was used for this site’s GFA.

• Toowong Shopping Village – was listed in 93 but absent from the 2016 directory. Web sources reveal the same retail GLA today, so the centre is deemed to have not been changed beyond cosmetic differences.

•

298

10.4. Appendix 4 – Details of documents analysed for plan performance

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Beenleigh Albert Shire Albert Planning No A compact urban centre Yes, In 2006 planning for Yes, Beenleigh came into Logan Council, Logan Scheme 1995; developed around specifically Beenleigh was administered specifically City in the 2008 council City Council Gold Coast predominantly urban area. under the 2003 Gold Coast amalgamations. The Gold Planning Scheme DCP1 covers Beenleigh. Planning Scheme. The Coast Planning scheme 2003; Logan Primarily focussed on RFGM is mentioned in the 2003 was used until its Planning Scheme commercial objectives, but planning intent for the incorporation into the new 2015 does include medium centre. Key objectives are Logan Planning Scheme density areas. to increase commercial and 2015. Listed as a principal residential growth in the activity centre, with intent centre statements aligned with SEQRP objectives. Plan notes at the beginning that SEQRP is incorporated. Beenleigh local plan specifically invokes regional policy

299

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Browns Logan City Logan Planning No Covered under "Browns Yes, The Logan Planning Yes, Listed as a major centre, Plains Council Scheme 1997; Plains Business Activity specifically Scheme 2006 was released specifically with intent statements Logan Planning Node District Strategy". in March 2006 and aligned with SEQRP Scheme 2006; Mixed industrial, retail, and incorporates the new objectives. Plan notes at Logan planning office centre. SEQRP. Centres are the beginning that SEQRP scheme 2015 specifically listed as being is incorporated. nominated in the SEQRP. No discussion of the actual principles of the SEQRP centre policy though.

Capalaba Redland City 1988 Town Yes - A draft Development Yes, Makes direct reference to Yes, Makes direct reference to Council Planning Scheme specifically Control Plan was released specifically the SEQRP. Desired specifically the SEQRP. Desired for the Shire of in 1996 but never adopted. Environmental Outcomes Environmental Outcomes Redland; Draft The draft strategic plan was specify that centres are to specify that centres are to Community also released in 1996, comply with SEQRP centre comply with SEQRP centre Concept Plan - which specifically principles. The centres are principles. The centres are Capalaba 1996; mentioned the RFGM and specifically listed in specifically listed in Redlands for this centre to be a Major conjunction with these conjunction with these Planning Scheme District Centre. This plan provisions. provisions. V1; Redlands was subsequently adopted Planning Scheme in 1998. The DCP primarily V7.1 focuses on business, entertainment, and culture uses, as well as streetscape improvements, but has little to say in terms of residential development.

300

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Carindale Brisbane City Town Plan for the Yes - Strategic plan describes Yes, Although the version of the Yes, The 2014 City Plan now Council City of Brisbane strategically support for RFGM through strategically 2000 scheme had been specifically specifically describes each 1987; Brisbane the pattern of development amended to July 2006, it centre in terms of its role in City Plan 2000; and establishing only references the RFGM. regional policy. Planning Brisbane City employment opportunities The RFGM is referenced scheme centre intents align Plan 2014 (3-8). The plan later only generally, in that the with overall regional acknowledges that the scheme aligns with the planning policy RFGM network has yet to intent of the RFGM. This is be finalised so the existing similar to the amended hierarchy centres is 1987 town plan. The maintained. This centre is a principles of the centre are "Regional Business Centre". aligned with general Strategic focus is one SEQRP and RFGM commercial uses rather principles of concentrations than residential of commercial uses and higher density residential.

301

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Chermside Brisbane City Town Plan for the Yes - Strategic plan describes Yes, Although the version of the Yes, The 2014 City Plan now Council City of Brisbane strategically support for RFGM through strategically 2000 scheme had been specifically specifically describes each 1987; Brisbane the pattern of development amended to July 2006, it centre in terms of its role in City Plan 2000; and establishing only references the RFGM. regional policy. Planning Brisbane City employment opportunities The RFGM is referenced scheme centre intents align Plan 2014 (3-8). The plan later only generally, in that the with overall regional acknowledges that the scheme aligns with the planning policy RFGM network has yet to intent of the RFGM. This is be finalised so the existing similar to the amended hierarchy centres is 1987 town plan. The maintained. This centre is a principles of the centre are "Regional Business Centre". aligned with general Strategic focus is one SEQRP and RFGM commercial uses rather principles of concentrations than residential. There is a of commercial uses and DCP for this centre which higher density residential. does not reference RFGM but already includes similar principles in terms of uses, including higher density residential uses.

302

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Cleveland Redland City 1988 Town Yes - The draft strategic plan Yes, Makes direct reference to Yes, Makes direct reference to Council Planning Scheme specifically released in 1996 specifically specifically the SEQRP. Desired specifically the SEQRP. Desired for the Shire of mentioned the RFGM and Environmental Outcomes Environmental Outcomes Redland; for this centre to be a Major specify that centres are to specify that centres are to Development District Centre. This plan comply with SEQRP centre comply with SEQRP centre Control Plan 2 - was subsequently adopted principles. The centres are principles. The centres are Cleveland 1991; in 1998. No changes to the specifically listed in specifically listed in Redlands existing DCP could be conjunction with these conjunction with these Planning Scheme identified in response to this provisions. provisions. V1; Redlands change. DCP discusses its Planning Scheme traditional "town" character, V7.1 and ongoing intent to support this and more residences. Highly focussed on design elements.

303

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Goodna Ipswich City Ipswich Strategic Yes - The plan discusses the Yes, The 2006 planning scheme Yes, Goodna is now specifically Council Plan 1989; strategically RFGM, however Goodna is strategically does not mention the specifically recognised as being an Ipswich Planning not specifically mentioned in SEQRP. Scheme continues activity centre under Scheme 1999; this context. Instead, it is to have some relationship regional policy, with aligned Ipswich Planning described as a major with the RFGM. Similar to strategic intent Scheme 2006; suburban centre and as a Ipswich CBD, it uses RFGM district centre in a hierarchy terminology of Goodna as a different to the RFGM. The "Major Centre". However 1989 strategic plan has the case is less clear as principles similar to RFGM other non-RFGM centres and allows potential for are also listed in this higher density residential category. The centre uses. This document hierarchy terms have suggests the Council changed since the previous prefers Redbank as the scheme however, to be preferred centre for eastern more in line with RFGM Ipswich. terminology. The RFGM is defined in the planning scheme dictionary, but it is not mentioned in the provisions relating to Goodna.

304

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Indooroopilly Brisbane City Town Plan for the Yes - Strategic plan describes Yes, Although the version of the Yes, The 2014 City Plan now Council City of Brisbane strategically support for RFGM through strategically 2000 scheme had been specifically specifically describes each 1987; Brisbane the pattern of development amended to July 2006, it centre in terms of its role in City Plan 2000; and establishing only references the RFGM. regional policy. Planning Brisbane City employment opportunities The RFGM is referenced scheme centre intents align Plan 2014 (3-8). The plan later only generally, in that the with overall regional acknowledges that the scheme aligns with the planning policy RFGM network has yet to intent of the RFGM. This is be finalised so the existing similar to the amended hierarchy centres is 1987 town plan. The maintained. This centre is a principles of the centre are "Regional Business Centre". aligned with general Strategic focus is one SEQRP and RFGM commercial uses rather principles of concentrations than residential of commercial uses and higher density residential.

305

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Ipswich Ipswich City Ipswich Strategic Yes - The Ipswich City Centre Yes, The 2006 planning scheme Yes, The consolidated Ipswich Council Plan 1989; Specifically Structure Plan from 1998 Specifically does not mention the Specifically planning scheme Ipswich Planning specifically mentions its SEQRP. Scheme continues recognises the Ipwich CBD Scheme 1999; RFGM status, and the key to have some relationship as a principal acticity Ipswich Planning principles from this plan. with the RFGM, as it does centre. Scheme 2006; These principles are describe the Ipswich CBD thoroughly incorporated into as a "key regional centre", the plan. Even before this the same terminology in the though, the strategic plan RFGM. The RFGM is already had similar defined in the planning principles in place. scheme dictionary, but it is not mentioned in the provisions relating to the Ipswich CBD.

Logan Logan City Logan Planning No Covered under "Business Yes, The Logan Planning Yes, Listed as a major centre, Central Council Scheme 1997; Activity Node District specifically Scheme 2006 was released Specifically with intent statements Logan Planning Strategy", with high in March 2006 and aligned with SEQRP Scheme 2006; employment activities, incorporates the new objectives. Plan notes at Logan planning retailing, office, medium and SEQRP. Centres are the beginning that SEQRP scheme 2015 high density res near specifically listed as being is incorporated. station. nominated in the SEQRP. No discussion of the actual principles of the SEQRP centre policy though.

306

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Logan Logan City Logan Planning No Covered under "Shailer Yes, The Logan Planning Yes, Listed as a major centre, Hyperdome Council Scheme 1997; Park Business Activity Node specifically Scheme 2006 was released Specifically with intent statements Logan Planning District Strategy". Primarily in March 2006 and aligned with SEQRP Scheme 2006; retail and office centre. incorporates the new objectives. Plan notes at Logan planning Evolved from a SEQRP. Centres are the beginning that SEQRP scheme 2015 Development Control Plan. specifically listed as being is incorporated. nominated in the SEQRP. No discussion of the actual principles of the SEQRP centre policy though.

307

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Mitchelton Brisbane City Town Plan for the Yes - Strategic plan describes Yes, Although the version of the Yes, The 2014 City Plan now Council City of Brisbane strategically support for RFGM through strategically 2000 scheme had been specifically specifically describes each 1987; Brisbane the pattern of development amended to July 2006, it centre in terms of its role in City Plan 2000; and establishing only references the RFGM. regional policy. Planning Brisbane City employment opportunities The RFGM is referenced scheme centre intents align Plan 2014 (3-8). The plan later only generally, in that the with overall regional acknowledges that the scheme aligns with the planning policy RFGM network has yet to intent of the RFGM. This is be finalised so the existing similar to the amended hierarchy centres is 1987 town plan. The maintained. This centre is a principles of the centre are "Major District Centre". aligned with general Strategic focus is one SEQRP and RFGM commercial uses rather principles of concentrations than residential of commercial uses and higher density residential.

308

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy North Lakes Pine Rivers Mango Hill Yes - Policies in the DCP plan are Yes, Policies in the DCP plan are Yes, This area is still covered by Shire Council, Infrastructure Specifically well aligned with the RFGM Specifically well aligned with the RFGM specifically DCP14 as of 2016. This Moreton Bay Development which is used to give which is used to give plan continues to reference Regional Control Plan 14 support to the proposal support to the proposal'. By the RFGM rather than the Council 1998; Moreton 2006, various precinct plans SEQRP. The 2016 Bay Regional had been prepared Planning scheme has its Council Planning generally in accordance entire strategic framework Scheme 2016 with the DCP aligned with SEQRP themes, so centres policy is well integrated with local policy. North Lakes is mentioned at the strategic level in the planning scheme

309

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Redcliffe Redcliffe City Redcliffe No Was not listed as activity No, but The planning scheme Yes, The old Redcliffe planning Council Consolidated centre in the RFGM SEQRP includes references to the specifically scheme was used by Planning Scheme documents or background mentioned SEQRP, but only in terms of Moreton Shire after the 1996; Redcliffe policies. No mention of subdivision outside the 2008 amalgamations, until City Planning RFGM, although the plan urban footprints. The the release of the new Scheme 2005; covers the peninsular area centres policy is not 2016 planning scheme. Moreton Bay and therefore includes mentioned. This implies The 2016 Planning scheme Regional Council some centre uses and RCC was aware of the has its entire strategic Planning Scheme medium density residential centre policy, but did not framework aligned with 2016 uses use it. SEQRP themes, so centres policy is well integrated with local policy. Redcliffe is mentioned at the strategic level in the planning scheme

Springfield Ipswich City Springfield Yes - Town centre policies in the Yes, Town centre policies in the Yes, The Town Centre Concept Council Structure Plan specifically Springfield structure plan Specifically Springfield structure plan specifically Plan has now been 1998; Ipswich are well aligned with the are well aligned with the updated to include Planning Scheme RFGM which is used to give RFGM which is used to give extended discussion on 1999; Ipswich support to the proposal' support to the proposal'. alignment the SEQRP Planning Scheme Curiously, the more detailed policies. The Springfield 2006; Springfield Town Centre Concept Plan structure plan in the Town Centre that guides development in Ipswich planning scheme Concept Plan the actual centre has no continues to specifically 2006; discussion on the SEQRP mention Springfield, and nor the RFGM. policy intent is reflected in these provisions

310

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Springwood Logan City Logan Planning No Covered under "Business Yes, The Logan Planning Yes, Listed as a principal activity Council Scheme 1997; Activity Node District specifically Scheme 2006 was released specifically centre, with intent Logan Planning Strategy". Primarily in March 2006 and statements aligned with Scheme 2006; industrial, retail, and office incorporates the new SEQRP objectives. Plan Logan planning centre. SEQRP. Centres are notes at the beginning that scheme 2015 specifically listed as being SEQRP is incorporated. nominated in the SEQRP. Beenleigh local plan No discussion of the actual specifically invokes principles of the SEQRP regional policy centre policy though.

Strathpine Pine Rivers Shire of Pine No Strathpine is considered as Yes, SEQRP is integrated into Yes, The 2016 Planning scheme Shire Council, Rivers Strategic the key centre in this part of specifically the plan, and Strathpine is specifically has its entire strategic Moreton Bay Plan 1988; the shire and the plan specifically listed as a major framework aligned with Regional Central Pine supports a range of centre activity centre. The strategic SEQRP themes, so centres Council Development type uses. purposes aligns with the policy is well integrated Control Plan No. SEQRP intent. with local policy. 8 1998; Pine Strathpine is mentioned at Rivers Plan 2006; the strategic level in the Moreton Bay planning scheme Regional Council Planning Scheme 2016

311

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Toombul Brisbane City Town Plan for the Yes - Strategic plan describes Yes, Although the version of the Yes, The 2014 City Plan now Council City of Brisbane strategically support for RFGM through strategically 2000 scheme had been specifically specifically describes each 1987; Brisbane the pattern of development amended to July 2006, it centre in terms of its role in City Plan 2000; and establishing only references the RFGM. regional policy. Planning Brisbane City employment opportunities The RFGM is referenced scheme centre intents align Plan 2014 (3-8). The plan later only generally, in that the with overall regional acknowledges that the scheme aligns with the planning policy RFGM network has yet to intent of the RFGM. This is be finalised so the existing similar to the amended hierarchy centres is 1987 town plan. maintained. This centre is a "Major District Centre". Strategic focus is one commercial uses rather than residential

312

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Toowong Brisbane City Town Plan for the Yes - Strategic plan describes Yes, Although the version of the Yes, The 2014 City Plan now Council City of Brisbane strategically support for RFGM through strategically 2000 scheme had been specifically specifically describes each 1987; Brisbane the pattern of development amended to July 2006, it centre in terms of its role in City Plan 2000; and establishing only references the RFGM. regional policy. Planning Brisbane City employment opportunities The RFGM is referenced scheme centre intents align Plan 2014 (3-8). The plan later only generally, in that the with overall regional acknowledges that the scheme aligns with the planning policy RFGM network has yet to intent of the RFGM. This is be finalised so the existing similar to the amended hierarchy centres is 1987 town plan. The maintained. This centre is a principles of the centre are "Major District Centre". aligned with general Strategic focus is one SEQRP and RFGM commercial uses rather principles of concentrations than residential of commercial uses and higher density residential.

313

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Upper Mount Brisbane City Town Plan for the Yes - Strategic plan describes Yes, Although the version of the Yes, The 2014 City Plan now Gravatt Council City of Brisbane strategically support for RFGM through strategically 2000 scheme had been specifically specifically describes each 1987; Brisbane the pattern of development amended to July 2006, it centre in terms of its role in City Plan 2000; and establishing only references the RFGM. regional policy. Planning Brisbane City employment opportunities The RFGM is referenced scheme centre intents align Plan 2014 (3-8). The plan later only generally, in that the with overall regional acknowledges that the scheme aligns with the planning policy RFGM network has yet to intent of the RFGM. This is be finalised so the existing similar to the amended hierarchy centres is 1987 town plan. The maintained. This centre is a principles of the centre are "Regional Business Centre". aligned with general Strategic focus is one SEQRP and RFGM commercial uses rather principles of concentrations than residential. There is a of commercial uses and DCP for this centre which higher density residential. does not reference RFGM but already includes similar principles in terms of uses, including medium density residential uses.

314

Centre Local Regulations 1996 1996 Centre notes 2006 2006 Centre notes 2016 2016 Centre notes government/s examined Reference Reference Reference to regional to regional to regional policy policy policy Wynnum Brisbane City Town Plan for the Yes - Strategic plan describes Yes, Although the version of the Yes, The 2014 City Plan now Central Council City of Brisbane strategically support for RFGM through strategically 2000 scheme had been specifically specifically describes each 1987; Brisbane the pattern of development amended to July 2006, it centre in terms of its role in City Plan 2000; and establishing only references the RFGM. regional policy. Planning Brisbane City employment opportunities The RFGM is referenced scheme centre intents align Plan 2014 (3-8). The plan later only generally, in that the with overall regional acknowledges that the scheme aligns with the planning policy RFGM network has yet to intent of the RFGM. This is be finalised so the existing similar to the amended hierarchy centres is 1987 town plan. Wynnum is maintained. This centre is a the only regional planning lower order District Centre. centre that is not listed with its own centre based local plan. It is distinctly residential in nature

315

10.5. Appendix 5 – Summed land area of DIS scores by centre and year

10.5.1. Residential DIS Table 71 - Summed land area (in hectares), by centre, of 2016 residential DIS

Locati Res Res Res Res Res Res Res Centre Name on Score 1 Score 2 Score 3 Score 4 Score 5 Score 6 Score 7 Beenleigh Outer 77 2 19 18 2 61 31 Browns Plains Outer 141 0 57 12 39 0 22 Capalaba Outer 102 0 20 10 22 36 4 Carindale Middle 93 0 91 0 17 0 16 Chermside Middle 172 0 44 0 12 22 80 Cleveland Outer 73 0 38 0 34 20 13 Goodna Outer 68 17 66 20 4 18 4 Indooroopilly Middle 76 0 77 0 24 22 22 Ipswich Outer 117 3 17 20 19 42 37 Logan Central Outer 47 0 96 33 29 27 13 Logan Hyperdome Outer 3 25 91 13 27 40 8 Mitchelton Middle 69 0 108 0 17 17 16 North Lakes Outer 124 0 25 22 5 36 31 Redcliffe Outer 65 0 125 4 18 20 1 Springfield Outer 103 0 398 0 0 0 112 Springwood Outer 55 2 27 0 1 8 80 Strathpine Outer 84 0 0 9 29 13 55 Toombul Middle 74 0 22 0 77 41 28 Toowong Inner 61 0 27 0 79 11 32 Upper Mount Gravatt Middle 51 0 61 4 1 16 67 Wynnum Central Middle 69 0 121 1 27 11 19

Table 72 - Summed land area (in hectares), by centre, of 2006 residential DIS

Locati Res Res Res Res Res Res Res Centre Name on Score 1 Score 2 Score 3 Score 4 Score 5 Score 6 Score 7 Beenleigh Outer 91 0 31 13 32 35 8 Browns Plains Outer 142 0 71 0 35 0 22 Capalaba Outer 124 0 6 3 22 36 4 Carindale Middle 94 0 93 3 5 7 16 Chermside Middle 136 0 46 37 64 14 33 Cleveland Outer 73 0 38 0 34 33 0 Goodna Outer 52 25 68 48 4 0 0 Indooroopilly Middle 71 0 77 1 49 23 0 Ipswich Outer 133 0 14 20 47 25 18 Logan Central Outer 37 0 128 31 33 10 7 Logan Hyperdome Outer 0 33 95 11 22 8 38 Mitchelton Middle 69 0 116 3 22 16 0 North Lakes Outer 108 0 35 25 0 48 28 Redcliffe Outer 81 0 119 12 14 6 1 Springfield Outer 155 0 396 0 0 0 62 Springwood Outer 55 0 57 19 18 0 26 Strathpine Outer 95 0 0 52 0 45 0 Toombul Middle 90 0 22 2 92 22 15 Toowong Inner 58 0 27 6 94 11 15

316

Upper Mount Gravatt Middle 50 0 71 6 11 21 40 Wynnum Central Middle 76 0 120 3 48 0 0

Table 73 - Summed land area (in hectares), by centre, of 1996 residential DIS

Locati Res Res Res Res Res Res Res Centre Name on Score 1 Score 2 Score 3 Score 4 Score 5 Score 6 Score 7 Beenleigh Outer 89 0 52 16 31 13 9 Browns Plains Outer 67 97 74 0 33 0 0 Capalaba Outer 114 24 13 18 25 0 0 Carindale Middle 94 0 95 12 17 0 0 Chermside Middle 182 2 48 15 46 0 37 Cleveland Outer 65 0 63 24 0 26 0 Goodna Outer 62 8 74 36 8 9 0 Indooroopilly Middle 77 0 77 26 41 0 0 Ipswich Outer 135 0 11 50 15 22 22 Logan Central Outer 71 0 120 0 54 0 0 Logan Hyperdome Outer 58 39 92 19 0 0 0 Mitchelton Middle 74 0 121 8 23 0 0 North Lakes Outer 60 42 38 16 0 88 0 Redcliffe Outer 94 0 120 12 6 0 0 Springfield Outer 427 0 96 0 0 91 0 Springwood Outer 64 0 73 0 36 0 0 Strathpine Outer 125 0 57 10 0 0 0 Toombul Middle 85 14 18 46 79 0 0 Toowong Inner 59 0 23 50 60 9 10 Upper Mount Gravatt Middle 50 0 78 5 5 50 12 Wynnum Central Middle 76 0 122 18 32 0 0

10.5.2. Commercial DIS Table 74 - Summed land area (in hectares), by centre, of 2016 commercial DIS

Locatio COM Score COM Score COM Score COM Score COM Score COM Score Centre Name n 1 2 3 4 5 6 Beenleigh Outer 6.0 174.0 21.5 0.0 5.6 3.6 Browns Plains Outer 1.6 207.7 21.7 1.4 16.1 22.1 Capalaba Outer 56.7 100.4 4.1 17.8 11.0 3.8 Carindale Middle 30.9 170.3 0.3 2.4 0.0 14.1 Chermside Middle 4.5 267.8 23.0 0.0 0.0 35.2 Cleveland Outer 50.4 105.5 5.6 0.0 4.1 12.4 Goodna Outer 27.2 128.5 34.5 0.1 0.0 7.4 Indooroopilly Middle 97.7 95.5 14.5 6.3 0.0 7.2 Ipswich Outer 41.7 146.2 8.4 53.5 0.0 6.2 Logan Central Outer 14.3 177.5 37.5 0.4 15.8 0.0 Logan Hyperdome Outer 112.5 30.9 15.4 0.0 39.8 8.4 Mitchelton Middle 44.4 158.4 8.5 0.0 0.0 14.8 North Lakes Outer 54.7 108.5 53.6 2.4 0.2 23.7 Redcliffe Outer 41.6 188.3 3.3 0.0 0.0 0.0 Springfield Outer 432.4 68.3 32.7 44.4 0.0 35.0 Springwood Outer 2.4 124.1 12.6 0.0 4.4 30.0 Strathpine Outer 5.9 112.0 17.0 0.0 0.0 56.4 Toombul Middle 43.2 172.2 2.8 4.4 0.0 20.2

317

Toowong Inner 33.8 160.5 1.9 3.2 0.0 11.3 Upper Mount Gravatt Middle 19.1 119.8 14.1 0.0 0.0 47.0 Wynnum Central Middle 101.1 132.1 1.9 3.3 0.0 9.2

COM Score COM Score COM Score COM Score Centre Name Location COM Score 1 2 3 4 5 Beenleigh Outer 40.0 137.8 17.2 7.4 8.4 Browns Plains Outer 70.8 152.9 11.2 1.4 15.5 Capalaba Outer 47.7 109.4 4.1 17.8 11.0 Carindale Middle 74.9 123.4 3.1 0.0 0.0 Chermside Middle 6.2 282.6 10.7 0.0 4.6 Cleveland Outer 50.4 105.5 5.6 0.0 16.5 Goodna Outer 10.8 169.1 0.1 9.8 7.9 Indooroopilly Middle 17.7 184.0 0.3 0.2 5.7 Ipswich Outer 37.2 153.7 21.4 23.4 14.0 Logan Central Outer 180.3 15.7 33.7 0.0 11.1 Logan Hyperdome Outer 139.4 0.0 15.4 0.0 14.3 Mitchelton Middle 11.6 194.5 3.0 0.0 3.6 North Lakes Outer 62.9 70.1 85.5 0.0 0.0 Redcliffe Outer 11.1 209.3 12.7 0.0 0.0 Springfield Outer 433.5 116.4 27.4 0.0 0.0 Springwood Outer 75.2 41.1 43.9 0.0 0.0 Strathpine Outer 70.6 64.1 13.0 43.8 0.0 Toombul Middle 33.8 186.8 4.4 0.0 1.9 Toowong Inner 17.0 182.9 0.9 0.0 3.6 Upper Mount Gravatt Middle 28.1 113.5 10.7 0.0 0.0 Wynnum Central Middle 3.7 229.7 2.8 11.3 0.0

Centre Name COM Score COM Score COM Score COM Score Location COM Score 1 2 3 4 5 Beenleigh Outer 86.4 85.8 16.3 7.9 13.4 Browns Plains Outer 0.6 189.9 56.0 0.0 24.1 Capalaba Outer 29.9 125.0 14.3 0.0 24.8 Carindale Middle 74.9 120.9 5.3 0.0 16.7 Chermside Middle 52.1 234.5 1.0 0.0 5.6 Cleveland Outer 95.4 46.3 22.0 3.7 10.6 Goodna Outer 52.2 121.4 16.0 8.1 0.0 Indooroopilly Middle 19.3 191.3 0.0 0.0 10.6 Ipswich Outer 63.8 105.0 67.4 14.9 0.0 Logan Central Outer 50.3 162.3 16.9 0.0 16.1 Logan Hyperdome Outer 13.6 143.2 9.1 0.0 41.1 Mitchelton Middle 11.4 192.7 4.4 0.0 17.7 North Lakes Outer 123.0 14.0 76.7 0.0 0.0 Redcliffe Outer 154.9 74.6 3.6 0.0 0.0 Springfield Outer 517.1 5.2 0.0 0.0 0.0 Springwood Outer 10.0 126.5 21.1 0.0 16.1 Strathpine Outer 97.2 38.0 29.4 26.8 0.0 Toombul Middle 17.9 207.7 0.0 0.0 17.2 Toowong Inner 17.0 181.7 0.0 0.0 12.0 Upper Mount Gravatt Middle 13.4 125.3 6.3 6.2 25.1 Wynnum Central Middle 3.7 229.1 1.9 0.0 12.8

318

10.5.3. Industrial DIS Table 75 - Summed land area (in hectares), by centre, of 2016 industrial DIS

Centre Name Location IND Score 1 IND Score 2 IND Score 3 IND Score 4 IND Score 5 Beenleigh Outer 168.9 41.8 0.0 0.0 0.0 Browns Plains Outer 142.2 51.5 0.0 34.8 42.2 Capalaba Outer 109.6 45.0 39.5 0.0 0.0 Carindale Middle 201.2 16.7 0.0 0.0 0.0 Chermside Middle 269.7 60.9 0.0 0.0 0.0 Cleveland Outer 139.7 19.4 18.9 0.0 0.0 Goodna Outer 151.5 24.2 14.6 7.4 0.0 Indooroopilly Middle 191.9 29.3 0.0 0.0 0.0 Ipswich Outer 198.0 11.6 30.7 15.7 0.0 Logan Central Outer 202.6 42.9 0.0 0.0 0.0 Logan Hyperdome Outer 143.4 58.2 0.0 5.4 0.0 Mitchelton Middle 199.9 26.2 0.0 0.0 0.0 North Lakes Outer 209.4 18.8 13.7 1.2 0.0 Redcliffe Outer 191.8 25.7 10.5 5.2 0.0 Springfield Outer 500.7 0.0 108.2 4.0 0.0 Springwood Outer 67.9 77.5 0.0 28.1 0.0 Strathpine Outer 103.9 61.1 0.0 17.5 8.9 Toombul Middle 197.3 43.6 0.0 1.9 0.0 Toowong Inner 193.4 16.2 0.0 1.1 0.0 Upper Mount Gravatt Middle 133.6 66.4 0.0 0.0 0.0 Wynnum Central Middle 228.1 12.9 0.0 6.5 0.0

Table 76 - Summed land area (in hectares), by centre, of 2006 industrial DIS

Centre Name Location IND Score 1 IND Score 2 IND Score 3 IND Score 4 IND Score 5 Beenleigh Outer 177.7 32.9 0.0 0.0 0.0 Browns Plains Outer 84.2 57.5 0.0 32.3 96.5 Capalaba Outer 110.6 43.9 37.0 0.0 2.5 Carindale Middle 198.3 19.6 0.0 0.0 0.0 Chermside Middle 290.1 40.5 0.0 0.0 0.0 Cleveland Outer 139.7 19.4 18.9 0.0 0.0 Goodna Outer 175.5 4.4 17.7 0.0 0.0 Indooroopilly Middle 199.7 20.8 0.0 0.8 0.0 Ipswich Outer 185.5 16.2 44.4 0.0 9.9 Logan Central Outer 203.2 33.6 8.8 0.0 0.0 Logan Hyperdome Outer 139.4 62.1 5.4 0.0 0.0 Mitchelton Middle 202.1 20.0 0.0 0.0 4.0 North Lakes Outer 190.9 30.4 0.0 21.8 0.0 Redcliffe Outer 204.2 14.3 14.6 0.0 0.0 Springfield Outer 557.8 35.7 0.0 19.5 0.0 Springwood Outer 92.7 43.6 13.6 10.7 12.9 Strathpine Outer 102.9 56.1 1.0 0.5 30.8 Toombul Middle 218.2 19.6 0.0 5.0 0.0 Toowong Inner 197.5 12.1 0.0 1.1 0.0 Upper Mount Gravatt Middle 135.4 58.4 0.0 6.2 0.0 Wynnum Central Middle 223.6 14.1 0.0 6.7 3.1

319

Table 77 - Summed land area (in hectares), by centre, of 1996 industrial DIS

Centre Name Location IND Score 1 IND Score 2 IND Score 3 IND Score 4 IND Score 5 Beenleigh Outer 126.1 84.6 0.0 0.0 0.0 Browns Plains Outer 18.5 106.9 97.1 48.1 0.0 Capalaba Outer 102.2 24.8 25.3 41.6 0.0 Carindale Middle 195.9 22.0 0.0 0.0 0.0 Chermside Middle 284.0 24.6 21.8 0.2 0.0 Cleveland Outer 141.1 29.4 7.1 0.4 0.0 Goodna Outer 165.4 25.4 6.9 0.0 0.0 Indooroopilly Middle 207.7 10.6 0.0 2.2 0.8 Ipswich Outer 198.1 12.2 45.6 0.0 0.0 Logan Central Outer 75.5 170.1 0.0 0.0 0.0 Logan Hyperdome Outer 59.2 146.1 0.0 1.6 0.0 Mitchelton Middle 200.1 22.1 0.0 0.0 4.0 North Lakes Outer 240.5 0.0 0.0 2.6 0.0 Redcliffe Outer 214.5 0.7 3.9 14.0 0.0 Springfield Outer 612.9 0.0 0.0 0.0 0.0 Springwood Outer 34.8 106.5 0.0 32.2 0.0 Strathpine Outer 110.2 31.4 35.4 0.0 14.4 Toombul Middle 206.0 17.2 13.9 4.8 0.8 Toowong Inner 197.5 12.2 0.0 0.1 0.9 Upper Mount Gravatt Middle 138.7 31.4 23.7 6.2 0.0 Wynnum Central Middle 222.2 14.7 0.0 7.3 3.3

10.5.4. Bulky goods retail Table 78 - Summed land area (in hectares), by centre, of 2016 bulky goods retail DIS

Centre Name Location BGR Score 1 BGR Score 2 BGR Score 3 Beenleigh Outer 180.0 21.5 9.2 Browns Plains Outer 187.0 10.7 72.9 Capalaba Outer 117.1 1.0 75.9 Carindale Middle 201.4 14.1 2.4 Chermside Middle 270.1 59.1 1.4 Cleveland Outer 142.6 0.0 35.4 Goodna Outer 151.2 24.5 22.0 Indooroopilly Middle 204.0 17.2 0.0 Ipswich Outer 202.7 7.6 45.7 Logan Central Outer 208.0 21.8 15.8 Logan Hyperdome Outer 153.4 0.0 53.6 Mitchelton Middle 211.7 14.5 0.0 North Lakes Outer 210.3 0.0 32.9 Redcliffe Outer 217.5 15.6 0.0 Springfield Outer 500.7 0.0 112.2 Springwood Outer 99.4 11.7 62.5 Strathpine Outer 112.8 1.5 77.1 Toombul Middle 216.0 26.8 0.0 Toowong Inner 195.6 15.1 0.0 Upper Mount Gravatt Middle 136.1 57.7 6.2 Wynnum Central Middle 235.4 12.1 0.0

320

Table 79 - Summed land area (in hectares), by centre, of 2006 bulky goods retail DIS

Centre Name Location BGR Score 1 BGR Score 2 BGR Score 3 Beenleigh Outer 177.7 0.0 32.9 Browns Plains Outer 184.9 1.4 84.3 Capalaba Outer 120.6 0.0 73.4 Carindale Middle 198.3 3.1 16.5 Chermside Middle 290.1 7.0 33.5 Cleveland Outer 142.6 0.0 35.4 Goodna Outer 175.6 4.4 17.6 Indooroopilly Middle 199.7 0.3 21.2 Ipswich Outer 190.7 11.6 53.6 Logan Central Outer 216.7 0.5 28.3 Logan Hyperdome Outer 149.4 0.0 57.6 Mitchelton Middle 202.1 3.0 21.0 North Lakes Outer 146.9 19.6 76.7 Redcliffe Outer 220.4 3.3 9.4 Springfield Outer 557.8 0.0 55.2 Springwood Outer 113.2 0.0 60.4 Strathpine Outer 109.2 0.0 81.7 Toombul Middle 219.0 1.8 22.0 Toowong Inner 197.5 0.9 12.2 Upper Mount Gravatt Middle 135.4 6.3 58.3 Wynnum Central Middle 223.6 2.8 21.1

Table 80 - Summed land area (in hectares), by centre, of 1996 bulky goods retail DIS

Centre Name Location BGR Score 1 BGR Score 2 BGR Score 3 Beenleigh Outer 187.8 12.4 10.4 Browns Plains Outer 188.1 0.0 82.6 Capalaba Outer 138.1 0.0 55.9 Carindale Middle 201.2 16.7 0.0 Chermside Middle 329.3 1.3 0.0 Cleveland Outer 141.1 18.8 18.1 Goodna Outer 182.5 8.2 6.9 Indooroopilly Middle 207.7 13.6 0.0 Ipswich Outer 190.2 15.3 50.4 Logan Central Outer 221.6 0.0 23.9 Logan Hyperdome Outer 152.5 3.3 51.2 Mitchelton Middle 204.5 21.7 0.0 North Lakes Outer 152.7 0.0 90.5 Redcliffe Outer 215.9 0.0 17.2 Springfield Outer 522.3 0.0 90.6 Springwood Outer 104.2 0.7 68.6 Strathpine Outer 151.3 0.0 40.0 Toombul Middle 219.9 22.8 0.0 Toowong Inner 197.7 12.9 0.0 Upper Mount Gravatt Middle 193.8 0.0 6.2 Wynnum Central Middle 224.1 23.4 0.0

321

10.6. Appendix 6 – Classification criteria for land use conformance

This appendix details the DIS conditions used to classify each land use into a type of conformance.

Table 81 - Classification criteria of conforming, under-developed, or exceeding uses

Detailed Use Conforming Under-developed Exceeding Vacant All = 1 If not conforming None Development Exclude Exclude Exclude Detached Dwelling Res = 3 or 2 Res > 3 Res = 1 Duplex Res = 3 or 2 Res > 3 Res = 1 Multi Unit Development Low Res >= 3 AND Res <=5 Res > 5 Res < 3 Multi Unit Development Mid Res >= 4 AND Res <=6 Res > 6 Res < 4 Multi Unit Development (Res = 5 AND Storey = 4) OR (Res = 6 AND Res > 6 AND Res < 6 AND storey > High Storey < 7) OR (Res = 7 and Storey > 6) Storey < 6 4 If not conforming Short term accommodation Com >= 2 None Warehouse IND >= 3 None If not conforming Manufacturing IND >= 3 None If not conforming Retail COM >= 2 None If not conforming Food and Drink Outlet COM >= 2 None If not conforming Bar COM >= 2 None If not conforming Office COM >= 2 None If not conforming Shopping Centre Strip COM >= 3 None If not conforming Shopping Centre Box COM >= 5 None If not conforming Bulky goods retail BGR >= 2 None If not conforming Health Care COM >= 2 None If not conforming Com <= 2, Res = 1 AND IND = 1 AND BGR = Library 1 If not conforming None Com <= 2, Res <= 4 AND IND = 1 AND BGR If not conforming Education facility = 1 None Com <= 2, Res <= 4 AND IND = 1 AND BGR If not conforming Hospital = 1 None Com <= 2, Res = 1 AND IND = 1 AND BGR = If not conforming Hall 1 None Com <= 2, Res = 1 AND IND = 1 AND BGR = If not conforming Support 1 None Com <= 2, Res <= 3 AND IND = 1 AND BGR If not conforming Place of worship = 1 None Com <= 2, Res = 1 AND IND = 1 AND BGR = If not conforming Open Space 1 None Indoor Entertainment Com >= 2 None If not conforming Public Infrastructure All = 1 If not conforming None Com <= 2, Res = 1 AND IND = 1 AND BGR = If not conforming Sport Complex 1 None Com <= 2, Res = 1 AND IND = 1 AND BGR = If not conforming Cultural Facility 1 None Service industries IND >= 2 None If not conforming Com <= 2, Res = 1 AND IND = 1 AND BGR = Community Centre 1 If not conforming None Main Street COM >= 3 None If not conforming Com <= 2, Res = 1 AND IND = 1 AND BGR = If not conforming Emergency Services 1 None Com <= 2, Res = 1 AND IND = 1 AND BGR = If not conforming Court House 1 None Car Park All = 1 If not conforming None Mixed Use Office Com > = 3 None If not conforming Mixed Use Industry IND >= 3 and COM >=2 None If not conforming Raw material processing IND >= 4 None If not conforming

322

Club Com >= 2 None If not conforming Mixed Use Residential Com >= 2 AND Res >= 4 None If not conforming Heavy Vehicle Parking All = 1 If not conforming None Res <= 2 AND COM <= 2 AND IND <= 3 Rural BGR = 1 If not conforming None Mixed Use Place of Worship Com >= 3 None All Com <= 2, Res = 1 AND IND = 1 AND BGR = Defence 1 If not conforming None Com <= 2, Res = 1 AND IND = 1 AND BGR = Cemetery 1 If not conforming None Mixed Use Complex Com >= 3 AND Res >= 4 None All Residential Complex RES >= 4 None All Commercial Unknown COM >=2 OR BGR >=2 OR IND >=2 None All

323

10.7. Appendix 7 - Correlation matrix of independent variables

Table 82 - Correlation matrix of independent variables

1996 1996 1996 1996

- - -

in residential in residential

199

Score

to 2016 to 2016 to 2016

Variable

1996 to 2016 1996

Unit Price 1996 Unit Price Band IEO 1996

1996 Employment Employment 1996

1996 Dwelling Mix Dwelling 1996

developed Sites in in Sites developed

1996 Density Score Density 1996 Under of Proportion

Change in industrial in industrial Change

1996 Compact Score Compact 1996 Change

Change in bulky good good in bulky Change

1996 Mixed use score use Mixed 1996

SNAMUTS Composite SNAMUTS

Change in commercial in commercial Change intensity zoning retail

zoning intensity intensity zoning intensity zoning intensity zoning Road Distance to CBD Distance Road

1996 Compact Score 1.000 1996 Density Score .863** 1.000 1996 Dwelling Mix .784** .802** 1.000 1996 Employment Score .475* 0.312 0.177 1.000 1996 Mixed use score .767** .496* .496* 0.211 1.000 Unit Price 1996 0.168 0.240 0.191 0.426 0.142 1.000 1996 IEO Band 0.286 0.274 0.109 .498* 0.263 .715** 1.000 SNAMUTS Composite 0.329 0.342 0.098 0.337 0.355 0.425 .814** 1.000 Road Distance to CBD -0.373 -.566* -0.322 -0.324 -0.243 -.530* -.743** -.736** 1.000 Change in residential zoning .645** .794** .662** -0.030 0.278 -0.140 -0.050 0.029 -0.251 1.000 intensity - 1996 to 2016 Change in commercial zoning intensity - 1996 to .474* 0.414 0.415 0.080 0.330 -0.074 0.259 0.227 -0.336 0.323 1.000 2016 Change in industrial zoning -0.326 -.531* -0.288 -0.319 -0.044 -0.306 -0.141 -0.093 0.414 -.533* 0.171 1.000 intensity - 1996 to 2016 Change in bulky good retail zoning intensity - 1996 to 0.155 0.187 0.124 0.124 0.035 0.135 0.301 0.196 -0.343 0.079 .670** 0.130 1.000 2016 Proportion of Under .625** 0.360 .575** 0.254 .770** 0.049 0.080 0.282 -0.111 0.172 0.215 -0.044 0.040 1.000 developed Sites in 1996 *. Correlation is significant at the 0.05 level (2-tailed), **. Correlation is significant at the 0.01 level (2-tailed).

324

10.8. Appendix 7 – Springwood summit event flyer

325

326

327

328

329