Planning for Green Open Space in Urbanising Landscapes

Planning for green open space in urbanising landscapes

Dr Christopher Ives, Dr Cathy Oke, Dr Benjamin Cooke, Dr Ascelin Gordon and Associate Professor Sarah Bekessy.

National Environment Research Program, Environmental Decisions Hub

School of Global, Urban and Social Studies RMIT University

Final Report for Australian Government Department of Environment

October 2014

© Christopher Ives, Cathy Oke, Benjamin Cooke, Ascelin Gordon and Sarah Bekessy Interdisciplinary Conservation Science Research Group School of Global, Urban and Social Studies RMIT University Melbourne VIC 3001 http://www.rmit.edu.au/socialhumanities/conservationscience/people

Copyright protects this material. Except as permitted by the Copyright Act, reproduction by any means (photocopying, electronic, mechanical, recording or otherwise), making available online, electronic transmission or other publication of this material is prohibited without the prior written permission of the Interdisciplinary Conservation Science Research Group.

Acknowledgements: Lake Macquarie City Council, Port Stephens Council, Meredith Lang Lower Hunter Councils, Survey methodology reviewers, Ailish Hehir, Luis Mata, Christopher Raymond, Amy Whitehead and residents who participated in our survey.

This report was funded by the Australian Government’s National Environment Research Program

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 2

1. Contents

List of Figures List of Tables Executive Summary COPYRIGHT PAGE ...... 2 Executive Summary ...... 6 1.1 Report Context ...... 7 1.2 NERP and the Lower Hunter Region ...... 8 1.3 Research Objectives ...... 8 1.4 Current Scientific Evidence ...... 10 2 Methodology ...... 15 2.1 Survey Design and Administration ...... 15 2.2 Map design and spatial analysis ...... 18 2.3 Statistical Analyses ...... 21 3 Results ...... 23 3.1 Survey Respondent Profile ...... 23 3.2 General Community Value Orientations and Perceptions of Green Space ...... 28 3.3 Community Satisfaction for Green Open Space ...... 33 3.4 General Green Open Space Values ...... 35 3.5 Mapping Values and Uses of Green Open Spaces ...... 36 3.6 Understanding People’s Favourite Green Open Spaces ...... 49 3.7 Factors Influencing Mapped Green Open Space Values (Suburb) ...... 51 3.7.1.2 Percentage Vegetation in Park ...... 55 3.8 Value Compatibility ...... 71 3.9 Qualitative Responses ...... 73 3.10 Summary of key findings ...... 74 4. Recommendations for Green Open Space Planning ...... 75 4.1 Recommendation One - Incorporate Values into Green Open Space Planning ...... 75 4.2 Recommendation Two - Consider Biodiversity Conservation Outcomes in All Green Open Space Planning Decisions ...... 78 4.3 Recommendation Three - Use Best Practice Green Open Space Planning Principles ...... 80 5. Application of Green Open Space Research to Strategic Assessment of Land Use Plans and Programs ...... 83 6 References ...... 85

Appendix A: Survey Booklet Appendix B: ABS 2011 census Appendix C: Ives et al (2014) Paper In Prep

List of Figures Figure 1. Age profile distribution by suburb ...... 24 Figure 2. Years Living in LGA, all respondents ...... 25 Figure 3. Dwelling Type ...... 25 Figure 4. Member of a Community Conservation Group, all respondents ...... 26 Figure 5. Income levels for all respondents ...... 27 Figure 6. Housing status, all suburbs, all respondents ...... 27 Figure 7. Community values for green open spaces, in general, for all respondents ...... 28 Figure 8. Importance of activities in green open spaces, all respondents ...... 29 Figure 9. Negative characteristics of green open spaces ...... 30 Figure 10. Satisfaction with amount of green open space, all respondents per suburb ...... 33 Figure 11. Satisfaction with the Quality of Green Open Space ...... 34 Figure 12. Accessibility of green open space ...... 34 Figure 13. Dot abundance for each attribute; LGA and suburb scale ...... 37 Figure 14. Dot abundance for all attributes, Nelson Bay ...... 41 Figure 15. Dot abundance for all attributes, Charlestown ...... 42 Figure 16. Dot abundance for all attributes, Toronto ...... 43 Figure 17. Dot abundance for all attributes, Raymond Terrace ...... 44 Figure 18. Dot abundance per 100m2 for nature and native biota values, and nature activities, Nelson Bay ...... 45 Figure 19. Dot abundance per 100m2 for nature and native biota values, and nature activities, Toronto ...... 46 Figure 20. Dot abundance per 100m2 for nature and native biota values, and nature activities, Raymond Terrace ...... 47 Figure 21. Dot abundance per 100m2 for nature and native biota values, and nature activities, Charlestown ...... 48 Figure 22. Factors related to favourite places in LGA ...... 50 Figure 23. Factors Related to Favourite Places in Suburb ...... 50 Figure 24 Park value dot abundance and park area per attribute ...... 53 Figure 24. (continued) Park value dot abundance and park area per attribute ...... 54 Figure 25. Park value dot abundance and percentage vegetation cover per attribute ...... 56 Figure 25. (continued) Park value dot abundance and percentage vegetation cover per attribute ...... 57 Figure 26. Park value dot abundance and park management category per attribute ...... 60 Figure 26. (continued) Park value dot abundance and park management category per attribute ...... 61 Figure 27. Park value dot abundance and distance to water per attribute (<500m from park) .. 63 Figure 27. (continued) Park value dot abundance and park management category per attribute ...... 64 Figure 28. Histogram all value dot abundance vs distance from place of residence ...... 65 Figure 29. Distance from place of residence plots for all attributes, by suburb ...... 67 Figure 30. Distance from place of residence plots for activity / physical exercise value for each suburb ...... 67 Figure 31. Distance from place of residence plots for social interaction value for each suburb .. 68 Figure 32. Distance from place of residence plots for nature value for each suburb ...... 68 Figure 33. Distance from place of residence plots for native biota value for each suburb ...... 69 Figure 34. Distance from place of residence plots for nature appreciate activities for each suburb ...... 69 Figure 35. Application of values in green open space decision making ...... 77 Figure 36. Conservation planning solutions for the Lower Hunter Region calculated for threatened species using the Zonation® package (image courtesy of Amy Whitehead)...... 79 Figure 37. Application of Green Open Space (GOS) Research Findings to Lower Hunter Regional Planning ...... 84

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 4

List of Tables Table 1 Highest Level of Formal Education ...... 26 Table 2 Main Occupation for all respondents ...... 26 Table 3 Spearman’s Rank Correlation Coefficient (Spearman’s Rho) ...... 31 Table 4 Relationships between socio-demographics and general values ...... 32 Table 5 Factor loadings for general open space benefits items, part 1 of the survey ...... 35 Table 6 Local Government Scale Map (Map 1) – Pearson’s R correlation coefficient ...... 39 Table 7 Suburb Scale Map (Map 2) – Pearson’s R correlation coefficient – Pearson’s R ...... 39 Table 9 Parameters for Zero-inflated Poisson models of park value dot abundance and percentage of vegetation in the park...... 55 Table 10 Parameters for Zero-inflated Poisson models of park value dot abundance and park management category ...... 58 Table 11 Parameters for Zero-inflated Poisson models of park value dot abundance and park distance from water ...... 62 Table 13 Value Compatibility Scores ...... 72

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 5

Executive Summary Green open space is considered key social and environmental infrastructure for a sustainable city. Good planning when space is limited or being planned for new urban areas is therefore especially crucial. Green open space planning is generally based on local policies and guidelines and is rarely informed by empirical evidence. Furthermore, very little research has been conducted on the variety of values that communities assign to green open space. There is therefore a need to explore empirically how ecological landscape features relate to people’s assigned values for these important community spaces. In this study we surveyed 418 residents of the Lower Hunter Valley region in NSW to explore how different values and activities are associated with various green open spaces. Map-based public participation GIS techniques were used to spatially define individual green open spaces and enable characterisation of participants’ surrounding landscape. The assignment of values was analysed spatially and related statistically to both physical characteristics of the park and socio-demographics.

People were found to assign a diversity of values to parks. Community values were especially strong at the local suburb scale compared to the Local Government Area (LGA) scale. Nature values and nature appreciation activities were identified as of similar importance as other benefits more typically associated with green open space, such as social interaction, health and recreation. Interestingly, the importance of nature values and activities was also high across all park management categories, including sports fields and general parks. Further analysis of the co-occurrence of value markers in individual parks revealed that values for native plants and animals were highly compatible with aesthetic, health, fitness values, but less compatible with social activities. Demographic factors were related to the strength and type of some values for green open space. In particular, age was related to native plants and animals values and health/therapeutic and cultural values. The configuration and design of green open spaces were also found to influence values. Park size was positively related to most value attributes, while intermediate levels of vegetation cover were related to many social and recreational activities. Additionally, parks closer to water bodies were considered more valuable for aesthetics and social activities. Whilst there was high satisfaction for green open spaces, it is important that negative perceptions of green open space are taken seriously as they can strongly reduce the value of green open spaces to communities, and reduce their use as a result.

Empirical data from studies such as this is important for the integration of community values into planning frameworks for green open space. These results suggest that many opportunities exist to maximise biodiversity outcomes in human-modified urban environments that are often overlooked in regional planning, while maintaining human landscape benefits and functions. We present a guide to the use of such data for green space planning in Australia, taking account of overlapping jurisdictional responsibility and existing guidelines and frameworks.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 6

1 Introduction

1.1 Report Context

Australia’s population is projected to grow significantly over the coming decades. A vital component of sustainable communities is the inclusion of green open space. This project seeks to build upon current best practice decision-making for the multiple environmental and social benefits of green open space at regional and local scales.

Part 1 of this research was a literature review of the multiple benefits of green open space, identifying the key environmental and social considerations for best practice planning.

We note that there are a variety of interpretations of green open space found in academic and grey literature including these found in the NSW Department of Planning Recreation and Open Space Planning Guidelines for Local Government (2010) for Open space: “publicly owned land that accommodates recreation facilities and provides spaces for recreational activities” and Urban public spaces: “publicly owned street and road reserves, lanes and town plazas and squares”.

Taking these and many other local government definitions review in our literature search we have a defined green open space as all publicly owned land that is set aside primarily for recreation, sports, nature conservation, passive outdoor enjoyment and public gatherings. This includes public parks, gardens, reserves, publicly owned forecourts and squares.

Part 2 of the project consisted of a questionnaire and survey of residents in the Lower Hunter region of NSW. Using maps this research investigated how the form and function of different types of green open space relate to community values and activities. This GIS methodology addresses theoretical gaps in the literature on spatial landscape values. Additionally, the data collected provides much more nuanced and detailed information than has been utilised in green open space planning previously, thereby addressing current gaps in best-practice planning approaches.

This report draws together findings from stages 1 and 2, with results from the literature review and empirical study synthesised into recommendations on how understanding values can contribute to green open space planning and design, targeted towards Australia’s future urbanising regions. These recommendations are applicable to the Australian Government’s

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 7

Sustainable Regional Development program and other relevant policies at federal, state and local levels.

1.2 NERP and the Lower Hunter Region

This project is one of many currently being conducted in the Lower Hunter Valley in NSW, funded by the Australian Government. Other projects include research on biodiversity conservation priorities in the region and assessment of community values. The research on community values in the Lower Hunter Region (Raymond & Curtis 2013) explored social values and development preferences for the region as a whole. The study focused on urban and rural residents as well as planning practitioners. By comparing areas of high social value with locations of development preference and conservation preference, areas of potential land-use conflict were identified.

Although adopting a similar methodological approach to Raymond and Curtis (2013), the present study was conducted at a finer spatial resolution. Further, the value attributes explored were altered to relate more closely to the issues relevant to local green open space planning. Green spaces provide significant environmental and social benefits within areas of urban development. Information on these will represent an important contribution to sustainability planning for the Lower Hunter Valley alongside the other research projects. More details about the National Environmental Research Program under which these projects fall can be found at http://www.environment.gov.au/science/nerp/research-hubs

1.3 Research Objectives

This research project is grounded in the theory that assigned values (the values individuals attached to physical places) are important in influencing environmentally significant behaviour (Seymour et al. 2010). Some research into people’s assigned values for green open spaces has already been conducted (e.g. Brown 2012), yet this area of study is still in its infancy and has not been pursued using rigorous public-participation GIS (PPGIS) methods in Australia. We have used proven methods for mapping assigned values (Brown 2005; Raymond & Brown 2006; Brown 2012) and activities (Brown and Weber 2011) to answer novel questions about green open spaces, how people interact with them and how this information can be related to conservation management decisions.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 8

In doing so the following key research questions have been pursued:

(1) What kinds of values and activities do people associate with green open space?

(2) Do general values for green open space held by people resemble those that are mapped spatially?

(3) How do values for green open space differ according to spatial scale?

(4) Which socio-demographic factors are most related to green open space values?

(5) Which environmental and landscape characteristics are related to mapped green open space values?

(6) Do green open space values reflect local government park management categories?

(7) How compatible are different green open space values?

(8) What opportunities exist for promoting conservation outcomes alongside social values for green open space?

We hypothesised that certain values and activities would be associated with different types of green open space, however these types of green open space may not align with previously defined categories according to current legal definitions under the Local Government Act 1993 No 30 [Parliament of NSW 1993). We also hypothesised that the values and activities identified by participants would differ between the suburb and Local Government Area (LGA) scale. For example values such as therapeutic value or children’s play activities could be most important within a close vicinity of residency while other recreation or biodiversity values are most important at the larger scale. While map scale has been found to influence the density of values assigned in similar mapping exercises (Brown 2012) and may affect their elicitation (Nielsen- Pincus 2011), there has been no research into these questions in the context of urban green spaces. It is this research gap that our study hopes to provide more details.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 9

1.4 Current Scientific Evidence

Green open space has been shown to be crucial to public health, personal wellbeing and is vital to the provision of urban ecosystem services and the maintenance of biodiversity in cities (Swanwick et al. 2003; CABE 2009; Healthy Spaces and Places 2009; Marshall and Corkery 2011; and Konijnendijk et al. 2013). These spaces offer city residents, workers and visitors benefits such as exercise, socializing, being in contact with nature and connecting with places of rich cultural heritage. The values people assign to green open spaces is reflected tangibly in higher prices of properties located in closer proximity to them (Crompton 2005; Cho et al. 2006; Sarev 2011). Further, they offer habitat for flora and fauna, along with other ecological benefits such as stormwater retention and management and urban heat island mitigation (Jorgensen and Gobster 2010; City of Boroondara 2011; Sports and Recreation Tasmania 2013). Below is a summary of the range of benefits of green open space as identified in Part 1 of this research project.

1.4.1 Benefits of Green Open Space

There has been a great deal of research on health and wellbeing benefits of green open space over the past few decades, providing evidence for these assertions. Studies have found that the provision of and access to green open space or gardens can be associated with longer life expectancy (Takano et al. 2002; de Vries et al. 2003; Giles-Corti et al. 2005), solace from stressful lives (Aspinall et al. 2013), better health or recovery from illness (Ulrich et al. 1991; Parsons et al. 1998), healthier weight range (Cummins and Fagg 2011; Lachowycz and Jones 2011), increased physical activity (Astell-Burt et al. 2013), greater socialisation (Germann- Chiari and Seeland 2004; Barbosa et al. 2007), increased community sense of place and belonging (Townsend and Weerasuriya 2010), and reduced levels of diabetes, heart disease (Astell-Burt et al. 2013 a and b) and depression (Townsend and Weerasuriya 2010). Some research suggests that health and wellbeing benefits of green open space are particularly pronounced for certain demographic groups such as the elderly (Takano et al. 2002), female adolescents (Boone-Heinonen et al. 2010) and youth (Maas et al. 2006). The importance of publicly accessible green spaces is also highlighted as especially important for those not physically or financially able to meet health objectives through other means (Maas et al. 2006; Byrne and Sipe 2010; Lee and Maheswaran 2010). Moreover, Wells (2000) found that children have higher self-worth and improved cognitive function as a result of contact with nature.

Many of the positive health benefits described above are not derived through physical access to green open space alone, but are influenced strongly by the quality of the space. The quality of

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 10

green open space is usually characterised by the diversity of facilities such as sporting fields or playgrounds, maintenance levels, shade availability, walking tracks, access to water, general amenity including safety (Broomhall 1996; Giles-Corti et al. 2005 Francis et al. 2012).

Using these characteristics of quality, Francis et al. (2012) found in their Perth, Australia, study that residents living in close proximity to high quality green open space were more likely to have low levels of psychological stress than those with low quality green open space, regardless of whether they used the green open space or not. Further, research has also shown that recovery from injury, surgery, or mental trauma is accelerated when patients are able to look at green open space, regardless of levels of physical access (Ulrich et al. 1991 and Byrne and Sipe 2010). Peschardt and Stigsdotter (2013) tested park characteristics that are related to stress relief in dense urban areas. The authors found that nature or the perception of nature in dense urban cities is important as a stress reliever. Studies by Ulrich et al. (1991), Grahn and Stigsdotter (2003), Nielsen and Hansen (2007), Korpela et al. (2010) and Ward Thompson et al. (2012) also found a positive correlation between access to green open space and reduced stress levels. Additionally, Aspinall et al. (2013) found green open space enhanced positive mood in busy city dwellers and Hu et al. (2008) found a relationship between the aesthetic quality (overall greenness) of public open spaces and the cardiometabolic health of urban residents and lower stroke mortality.

An obvious health benefit derived from access to green open space is the opportunity for physical activity, both through formal or informal recreation. Indeed, for certain activities the size and amount of green open space available is just as important as the quality of the space (Giles-Corti et al. 2005). In an Australian study, interview respondents with good access to large and attractive parks were found to be twice as likely to engage in physical activity in public green open space (South Australian Active Living Coalition. 2010). Likewise, residents in Perth reported that they were 50% more likely to increase their walking if green open space was large and attractive (Commonwealth of Australia 2010), and Boone-Heinonen et al. (2010) found that availability of parks led to increased participation rate in active sports, especially for female adolescents. Brown et al. (2014) also observed a positive linear correlation between the size of urban parks and levels of physical activity in an Adelaide study, although this was stronger for certain park types than others. Finally, connectivity of green open space is extremely important for its utilisation (Giles-Corti et al. 2005). United States and European research has shown that green open spaces with connection to other green open space such as bike or walking paths increase physical activities and length of time spent at the park (see Byrne and Sipe 2010 and references therein).

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 11

In addition to the numerous physical and mental benefits mentioned above, there are many benefits humans derive from the ecological functions green open spaces perform. City residents benefit from temperature regulation and reduction in urban heat island effect (Memon et al. 2008; Smardon 1988), reduction in wind capacity (Young 2011; Bernatzky 1982), filtering of air (Bernatzky 1982), reduction in city noise (Fang and Ling 2003; Gidlöf-Gunnarsson and Öhrström 2007), sequestering carbon (Tratalos et al. 2007), and stormwater attenuation and reduction in flooding through stormwater retention and water sensitive urban design and reduce reflected light (Young 2011; Samuels et al. 2010). Many of these benefits can provide ‘insurance’ against changes in future environmental conditions such as extreme temperatures floods or storms, and provide a strong case for investment by planning authorities (Commonwealth of Australia 2010; City of Melbourne 2012; Byrne and Sipe 2010).

Green open spaces such as parks, reserves and corridors are known to be significant ecological features that provide for the maintenance of species populations and ecological function within urban landscapes (Savard 2000; Sandstrom et al., 2006). In short, three factors are known to greatly influence the contribution of green open space to the biodiversity of urban landscapes: geometry (size and shape), landscape configuration and habitat structure.

First, it is generally considered that larger reserves harbour greater levels of biodiversity than smaller reserves (Cornelis and Hermy 2004). This has been found for a number of taxa (see Drinnan 2005) including arthropods (Gibb and Hochuli 2002), but is especially critical for birds, which usually have large home ranges (Donnelly and Marzluff 2004).

Second, the configuration of green open spaces throughout an urban landscape influences the biodiversity within them as a function of ecological connectivity and the effect of the developed areas between reserves (the urban matrix). Corridors (linear vegetation strips) have potential to serve as habitat linkages and increase the flow of organisms and their genes (Bolger et al. 2001). However, it is known that human activities and urban structures outside of green open spaces affect the composition of species because of changes in environmental conditions. This is most clearly seen in differences in species composition along a gradient of high to low urban intensity (often referred to as the urban-rural gradient) (McDonnell et al. 1997; McKinney 2008). For this reason, scholars have emphasized the importance of improving landscape connectivity by considering the contribution of private green spaces such as gardens and backyards to biodiversity conservation (Rudd et al. 2002; Goddard 2010).

Finally, the biodiversity that is present within green open spaces depends greatly on the habitat quality and structure within them. In many ways, this is the fundamental consideration since

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 12

reserve size and connectivity are only important if potential habitat exists within a green open space in the first place. While some species are resilient and adapted to harsh urban conditions, many require particular habitat characteristics. The complexity of habitat (often defined by vegetation structure) within green open spaces (including reserves) has been shown to influence the species composition of birds (Sandstrom et al., 2006; Donnelly and Marzluff 2006), vertebrates (Garden et al., 2007) and arthropods (Lassau and Hochuli, 2004). Further, the presence of exotic species can affect the biodiversity of urban reserves because of resulting physical habitat transformation or interspecific competition (see for example, Ives et al. 2013).

An expanded discussion on the benefits of green open spaces in an urban environment can be found in the full literature review (part 1) of this research project.

1.4.2 Social Values and Green Open Space

Given the range of benefits green open space provides, understanding how different groups within a community use these spaces is highly valuable as is knowledge of how these groups may change their use in the future. A complementary set of information that should also be considered when designing a green open space network is that of community values. The values that are of most relevance to green open space planning are those that people have for particular places. These are known in the scientific literature as “assigned values” and are generally defined as “the expressed relative importance or worth of an object to an individual or group in a given context” (Brown 1984). Values are fundamental in influencing people’s attitudes (preferences) towards particular spaces and the benefits they receive from them. For this reason, the values people assign to places are also an important component of landscape planning (Ives & Kendal 2013; Swanwick 2009). Understanding the compatibility of different values and uses of green spaces can be critical to designing landscapes that fulfil the needs of local populations and minimise conflict (Brown & Reed 2012). In the case of green open space, values may include physical activity, aesthetics, cultural heritage, social interaction etc. with each of these values being weighted more or less greatly depending on the person.

Green open space values are important to study as they are intimately linked with wellbeing and sustainability, yet are not as easily measured as visitation behaviour. However, park visitation can be a coarse measure of the benefits people derive from the park since people may visit parks for various reasons. Measuring community values allows people to express their satisfaction with parks and green open spaces for different reasons and provides an indication of how adequately the parks are meeting their needs. Moreover, values are particularly important in the context of environmental protection as people may highly value certain

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 13

functions of the park (such as biodiversity protection, air filtration, healthy water) without visiting the park or receiving direct benefit from it. Values are thus a useful measure when seeking to understand the interaction between people and green spaces. Accordingly, a number of research studies have emerged that have explored park values in a spatial manner (see Tyrvainen et al. 2007; Tyrvainen et al. 2003; Brown 2008).

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 14

2 Methodology

2.1 Survey Design and Administration

This project utilised Public Participation GIS (PPGIS) methods adapted from previous research (Brown 2005; Raymond & Brown 2006; Brown 2012; Brown and Weber 2011) to examine the distribution and types of values and activities of residents of the Lake Macquarie and Port Stephens Local Government Areas (LGAs). The mail out included a questionnaire (survey) designed to answer the research objectives previously outlined, as well as a mapping component to identify spatially the distribution of values and activities. This technique was modelled on the principles outlined in Brown (2005) and Raymond and Brown (2006).

The survey questionnaire booklet, maps and sticker dot legend components of the mail out can be seen in the Appendix A.

Specifically, the survey instrument was comprised of six sections:

(1) General benefits people associate with green open space (this did not reference any specific locations)

(2) Mapping green space values, activities and negative qualities at both (i) LGA and (ii) suburb scales. This involved placing mnemonic sticker dots from a ‘map legend’ onto a paper map with 2 panels.

(3) Qualities of people’s favourite places at the LGA scale

(4) Qualities of people’s favourite places at the suburb scale

(5) Socio-demographic questions about the survey respondents, and

(6) Other comments about general levels of satisfaction with the local green space network.

Parts (1) and (2) of the survey instrument (general values and mapped values) contained identical phrases. This enabled similarities and differences to be explored between how people value green open space in a general sense, and how they value specific green spaces in their local area. The question items used were developed with reference to values studied in other PPGIS studies (e.g. Tyrvainen et al., 2007; Brown, 2005). Similarly, the types of attributes related to people’s favourite places at the suburb and LGA scale were modified from those used by Jim and Chen (2006).The final typology of values was selected after consultation with academic colleagues, local government planners and practitioners in the Lower Hunter, and two focus groups with the public in the Lower Hunter. This process also highlighted areas where phrasing of questions could be clarified to enhance the construct validity of the survey.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 15

Survey packets were mailed to 1000 residents within the Lower Hunter region. Participants were selected from four suburbs within Port Stephens and Lake Macquarie LGAs. These were Raymond Terrace, Nelson Bay (Port Stephens), Charlestown and Toronto (Lake Macquarie). These areas were chosen because they offer contrasting biophysical and socio-demographic profiles. The LGA was chosen as the appropriate scale of analysis (as opposed to the entire Lower Hunter area) because open spaces as urban green spaces are managed by local government and are generally engaged with by local communities. Due to the survey design requiring participants to provide information about green spaces at two different scales, it was necessary to select participants from individual suburbs nested within the wider LGA.

Participants were selected from spatially-referenced residential households to enable stratification by suburb. Residential address data was obtained by a market research company, Truscott Research, using a database of mobile phone records and non identifiable geospatial information for properties under licence from PSMA via the Australian Government Department of Sustainability, Environment, Water, Population and Communities (SEWPaC), now Department of Environment (DoE). The dataset used by Truscott Research contains residential addresses and telephone numbers only and no health, personal or sensitive information and was used for the sole purpose of the research project, and solely to obtain consent to send the survey, which was sent back by individuals, and to link (via a non identifiable code) their responses to the spatial file provided by DoE. Truscott Research abides by the National Privacy Principles (“NPPs”) under the Privacy Act 1988 (Cth) (“Privacy Act”) and the Privacy Amendment (Private Sector) Act 2000. As members of the Australian Market and Social Research Society, Truscott Research abides by the principles of the AMSRS Code of Professional Behaviour.

Upon receiving the call, the householder was asked to indicate their willingness to participate in the research project, their name and address. Consent was therefore established verbally. Participants were asked to indicate their age according to three brackets: 18-35, 35-55, >55. This information was used to stratify participants as follows: >20% 18-35, >20% 35-55. This was deemed advisable given the bias towards older respondents in these kinds of surveys.

The survey mail out, which included a consent form, was then sent to individuals who provided verbal consent to do so. This telephone recruitment technique was established from earlier research in the Lower Hunter by Dr Christopher Raymond (Charles Sturt University), and found

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 16

to be effective for ensuring adequate survey response rates. The Dillman method was also employed to maximise survey response rates (Dillman 2007). This consists of an initial survey mail out, inclusion of a gift (in this case 6 packaged postal stamps) reminder postcards and resending of new survey packets to non-respondents. Based on previous similar research (Brown 2005) a 50% response was expected.

In part 1 of the survey residents were ask their view as per a scale of 1-5 to a series of questions to understand the benefits they gained from green open spaces in general. The same values, activities and characteristics were used for the mapping exercise for specific parks in part 2. The full survey is found in Appendix A, however we have included the three components of part 1 below so it can be easily referred to as part of the results section.

Not at all A little Somewhat A lot A great deal 1 2 3 4 5

Question 1 “How much do you value the following aspects of green open space in general?” Values: 1: Aesthetic / Scenic (i.e. the visual attractiveness of a place) 2: Activity / Physical Exercise (i.e. opportunities for physical activity) 3: Native Plants and Animals (i.e. the protection of native biodiversity) 4: Nature (i.e. experiencing the natural world) 5: Cultural Significance (e.g. appreciating culture or cultural practices such as art, music, history and indigenous traditions) 6: Health / Therapeutic (i.e. mental or physical restoration) 7: Social Interaction (i.e. opportunities to interact with other people)

Question 2 “In general, how important to you are the following activities undertaken in green open spaces? Activities: 8: Casual recreation (e.g. walking, kite flying, throwing Frisbee, walking dog etc.) 9. Exercise for fitness (e.g. jogging, cycling, walking, group sports) 10. Social activities (e.g. picnics, barbeques etc.) 11. Children’s play (e.g. areas for children to explore, have fun etc.) 12. Nature appreciation (e.g. bird watching, bush walking, photography etc.)

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 17

Question 3 “How much would the following characteristics of a green open space reduce its value to you? Characteristics: 13. Unappealing (e.g. neglected, damaged, anaesthetic, ugly) 14. Scary/Unsafe (e.g. dangerous or threatening) 15. Noisy (e.g. disturbingly loud or noisy) 16. Unpleasant (e.g. too hot, too windy, no shade or shelter etc.)

The numbers 1 – 16 above against the values, activities and characteristics are used in the results section for these attributes

2.2 Map design and spatial analysis

Map design considerations included the need to balance competing objectives between spatial accuracy and collecting accurate data. For example:

• In order to get a good spatial accuracy in the responses, the map needed be large and provide ample information and reference points to allow survey participants to easily orient themselves and identify relevant green open spaces.

• To encourage a good response rate by not deterring or overwhelming participants it was important to keep the map clear and easy to read.

• The value stickers had to be small for spatial accuracy but large enough that survey participants could easily peel and place the stickers on the map

• Printing scanning and general user friendliness restricted the size of the maps to A1.

Numerous spatial data sources were used to generate maps and analyse data. The key data sources were as follows: • Road features (labelled) from respective council, Lake Macquarie City Council, © Lake Macquarie City Council, 2013, Port Stephens City Council, © Port Stephens City Council, 2013, • Local Park Layers (labelled) from respective council data, Lake Macquarie City Council, © Lake Macquarie City Council, 2013, Port Stephens City Council, © Port Stephens City Council, 2013, • National Parks

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 18

Department of Sustainability, Environment, water Population and Communities, © Copyright of Commonwealth of Australia, 2012, • State Parks Land and Property Information, © Copyright of NSW Government, 2012, • Vegetation layer Hunter Councils Inc., © Hunter Councils Inc., 2005, 20m resolution Raster (Wooded Vegetation) • Water body Layer Hunter Councils Inc., © Hunter Councils Inc., 2005

Following receipt of the completed surveys and maps, the location of sticker dots were digitised into a Geographic Information System (GIS). Each sticker dot was assigned to a respondent’s survey ID and the value attribute was recorded. Invalid survey points (points that did not adhere to the survey instructions) were digitised but omitted from analysis. This task was completed by a qualified GIS technician. In addition to generating a digital spatial point dataset of mapped values, some processing and editing of other data layers was necessary to enable later analysis. These are outlined below.

Refining polygon geometry

Whilst the original park layers where adequate for visualising the Green Open Spaces in the maps sent to Survey Participants, the polygons in the original layers were not appropriate for analysis. The park polygons were edited to more accurately represent open green spaces while preserving the size and geometry of areas of continuous character. Park polygons that shared a boundary and the same park character were merged into a single park polygon using the Dissolve tool. The layer was manually inspected using a combination of Google Earth, satellite imagery, and a local street directory1 as validation data. Further edits were made according to the following rules/principles:

• Where an obvious change in physical character was apparent from satellite imagery the polygon was split and the character edited

• Polygon boundaries were edited to and reflect their true size.

• Adjoining local council parks were merged if all of the following criteria were satisfied; maximum 30m separation, vegetation coverage, land use unchanged and not interrupted by any roads or linear features other than streams and paths

1 Gregory’s Newcastle Street Directory, 28th Edition FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 19

• Omitted green open spaces were added if they were identified as a park in one of the validation references and the area was accessible to the public

• For National Parks, tracks and unsealed roads were not deemed to interrupt a person’s use or appreciation of the area. Therefore, where these features altered the park shape, the polygon was edited so the park area and perimeter reflected the park geometry without being distorted by other features.

The XY coordinates of each open space polygon centroid was calculated to aid in proximity calculations later on. In addition, the minimum distance of each park boundary to a water body was calculated.

Respondent home address

The address or nearest street corner of the each respondent (from questionnaire response) was digitised in a separate point feature class. The SurveyID field was stored with the “Home” point and the X and Y coordinates of the Home Point objects were added as fields in the attribute table (using the Add XY tool in ArcGIS). ArcGIS tools were used to relate these locations to other relevant information:

• The distance (as the crow flies) to the nearest park to their home residence, generated by the ArcGIS tool “Near”. This was calculated as:

�� ! =

( Home_Xm − Park_Center_Xm ! + (Home_Ym − Park_Center_Ym )!)

• The percentage of vegetation cover within a 100m radius of respondents’ home residence calculated using the vegetation layer and the “Spatial Join” tool.

Park Management Categories

To ensure consistency, we reclassified the categories parks were assigned to because the two LGAs that supplied park data did not use the same classification system within their local planning documents.

Parks from council data layers were reclassified as:

• Sportsfield – an area designated for sports (i.e. ovals, golf courses, etc.)

• General – the park was dominated by a designated community function (i.e. children’s parks, landscaped green open spaces)

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 20

• Natural – an area that has generally naturally occurring vegetation. Area is devoid of obvious evidence of human interference with vegetation and landscape except for grass length management (mowing). This category also included National and State Parks.

In the Lake Macquarie LGA Public Parks and General Community park management categories were reclassified as General, however, Natural Area and Sportsfield were mostly unchanged. For Port Stephens, the original classes Urban Park, General Community, Foreshore and Cultural Significance were merged into the new class General. Natural Area and Sportsfield classes did not require any change.

As school fields are an open space often used by sporting clubs and other community groups the School yards were added to the Green Open Spaces polygon layer for analysis, classed as school these features were manually added in from a Council provided list, the polygons were created to represent the Green Open Space portion of school grounds.

Mapping point attribute densities To display the digitised data, ‘heat maps’ for each value attribute were produced as follows: • A base polygon layer consisting of 100m2 polygons that spanned the extent of the Suburb scale was created using ArcMap • Survey Point layers were generated for each value attribute type. • Each Value Attribute Point layer was spatially joined to the base layer. Each Spatial Join generates a “Join_Count” field which is the amount of points that intersect the 100m square polygon • The resulting density maps were then visualised according to the density of points in each 100m2 grid.

2.3 Statistical Analyses A range of statistical methods was employed to analyse data from the survey and spatial mapping responses. All analysis was conducted using the R statistical environment (R Core Team 2014, vers 3.1.0). Details of the main analyses are outlined below:

Relationships among ordinal Likert scale survey responses and between other survey responses were tested via Spearman Rank Correlation analysis. Pearson correlation tests were performed between continuous variables. Differences between categorical factors (e.g. housing status) were analysed via Chi-squared tests. Survey responses from Parts 1 and 5 of the survey (general

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 21

open space values and socio-demographics) were also related to the abundance of mapped value dots via correlation analyses.

To simplify the range of green space values and identify general themes in how people relate to green spaces, a factor analysis was conducted on responses from Part 1 of the survey instrument. First, a scree plot of the data was generated from a principal components analysis to determine the number of distinct factors. The factor analysis was then conducted using the “factanal” function in the “stats” package in R using varimax rotation.

Relating the abundance of mapped value dots in parks to landscape and environmental variables required a statistical method that accounted for the fact that a high proportion of the parks in the study regions did not contain any dots. Since the response variable (dots in parks) was count data, zero-inflated Poisson modelling was adopted for this analysis. These analyses were conducted using the ‘zeroinfl’ function in the ‘pscl’ package in R. Before conducting the analysis, all predictor variables were standardised by subtracting the mean of the dataset from each value and dividing by the standard deviation.

To analyse the effect of distance from home on the assignment of value dots, it was necessary to account for the configuration of parks in each suburb in order for the true effect to be ascertained. To this end, a null model of park values was generated by randomly assigning 6 ‘dots’ to parks in each suburb for each respondent’s home address. The resulting output represented a random distribution of park distances from home addresses which could then be compared to the real mapped data. Histograms were produced for each value attribute for both the null models and real datasets. The differences in the bin values of each histogram were then plotted as a way of representing the true effect of distance from home on park values.

The compatibility between mapped values was calculated by comparing dot abundances in park polygons. A value compatibility score between value attributes V1 and V2 in a park was calculated as follows:

Value compatibility score (V1, V2) = 1 – | (V1 – V2) / (V1 + V2) | This gives a compatibility score (ratio) between the pair of value dots. The mean score for each value pair was calculated by averaging over all parks. A matrix of pairwise value comparisons was generated by repeating for every value pair.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 22

3 Results

3.1 Survey Respondent Profile

The response to our survey was 418 completed questionnaires and maps out of a possible 972 (28 of the 1000 that were sent out were Return to Sender), which equates to a response rate of 43%. When broken down by suburb the responses were: Raymond Terrace, 77; Nelson Bay, 121; Charlestown, 116; and Toronto, 104.

50.6% of respondents were male and 43.3% female. 93% of respondents nominated the contact address as their principle place of residence.

The age profile distribution of our respondents by suburb is shown in Figure 1 below. The median respondent age for the four suburbs were as follows: Charlestown – 62; Toronto – 61; Nelson Bay – 60.5; and Raymond Terrace – 57.

Although the respondents were biased towards an older demographic compared to 2011 census data (see Appendix B for summary), there is sufficient variability in the dataset to explore the values and preferences of younger people. Further, we note that this project does not aim to survey a representative sample of residents from these locations as a way of providing data to directly inform the planning of specific suburbs. Instead, we have sought to use this data to draw out general trends in how green space is valued that can be applied to urban and regional planning more broadly.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 23

Charlestown Toronto

Nelson Bay Raymond Terrace

Figure 1. Age profile distribution by suburb

Sixty five respondents had children under 9 years old (with a mean of 1.9 children), and fifty eight people had children between 10 and 17 years old (with a mean of 1.5 children). Figure 2 shows the number of years respondents had lived in their LGA, with the highest proportion of residents have been living in the region for approximately 10 years, although a substantial number have been in the area much longer.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 24

0.020 0.015 0.010 Proportion of Respondents 0.005 0.000 0 20 40 60 80 100 Years Living in LGA

Figure 2. Years Living in LGA, all respondents

Figure 3 shows that the vast majority of respondents lived in detached houses, compared with other forms of dwelling types.

80 House Townhouse Flat Other 60 40 20 Number of Survey Respondents Number of Survey 0 Charlestown Nelson Bay Raymond Terrace Toronto

Figure 3. Dwelling Type

The number of people engaged in a community conservation group is shown in figure 4. Engagement was low for all suburbs, but particularly for Raymond Terrace.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 25

100 No Yes 80 60 40 20 0 Charlestown Nelson Bay Raymond Terrace Toronto

Figure 4. Member of a Community Conservation Group, all respondents

Table 1 shows that the majority of respondents have tertiary education qualifications.

Table 1 Highest Level of Formal Education Highest Level of Formal Education Count University or Technical Institution 149 Technical or Further Education Institution 125 Secondary School 109 NA 27 Primary School 7 No formal schooling 0

Retirees dominate the main occupation response profile (table 2), reflecting the bias towards an older demographic. Although differing from census results (Appendix B), this is not unexpected given the type of survey administered, and the composition of some of the suburbs (particularly Nelson Bay).

Table 2 Main Occupation for all respondents Occupation Count Retired 178 Other 56 Professional 46 Home duties/parenting 25 NA 25 Manager 23 Clerical or administrative worker/ Sales worker 18 Technician or trades worker 18 Community or personal service worker 13 Student 12 Machinery operator or driver 3 Farmer 0

Figure 5 shows that there was a reasonable mix of income levels in all suburbs.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 26

15 Negative Nil $0−$10,399 $10,400−$15,599 $15,600−$20,799 10 $20,800−$31,199 $31,200−$41,599 $41,600−$51,999 $52,000−$64,999 $65,000−$77,999 5 $78,000−$103,999 $104,000+ Number of Survey Respondents Number of Survey 0 Charlestown Nelson Bay Raymond Terrace Toronto

Figure 5. Income levels for all respondents

More people owned their own home (figure 6) than any other housing status for all suburbs. However, the proportion of mortgage holders was higher in Raymond Terrace than other suburbs.

Own Outright 60 Own Mortgage Rent

50 Public Housing Other 40 30 20 Number of Survey Respondents Number of Survey 10 0 Charlestown Nelson Bay Raymond Terrace Toronto

Figure 6. Housing status, all suburbs, all respondents

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 27

3.2 General Community Value Orientations and Perceptions of Green Space

This section details responses from Part 1 of the survey instrument, which explored the values, important activities and negative qualities of green open space in a general sense. The scale of 1-5 used in the figures below refer to responses to questions regarding the benefits gained from green open spaces in general. 1 = Not at all; 2 = A little; 3 = Somewhat; 4= A lot; and 5= A great deal.

3.2.1 Community Values for Green Open Spaces

All values for green open space rated highly amongst respondents, as can be seen in Figure 7 below. Cultural significance was the lowest rated attribute, but all values had means of above 3. Of particular interest to biodiversity conservation is the fact that ‘native plants and animals’ and ‘nature’ rated higher than ‘social interaction’ and ‘health and therapeutic value’.

2 Mean Response

0 Nature Aesthetic/Scenic Social Interaction Health/Therapeutic Cultural Significance Cultural Activity/Physical Exercise Activity/Physical Native Plants and Animals Native

Figure 7. Community values for green open spaces, in general, for all respondents

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 28

3.2.2 Importance of Activities in Green Open Spaces

Items about important activities were all rated highly, see Figure 8 (mean response approximately 4). In a similar way to green space values, nature appreciation activities were rated as highly as other activities more traditionally considered in green open space planning (e.g. casual recreation and exercise for fitness).

2 Mean Response

0 Childrens Play Social Activities Casual Recreation Nature Appreciation Exercise For Fitness For Exercise

Figure 8. Importance of activities in green open spaces, all respondents

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 29

3.2.3 Negative characteristics of green open spaces

All four negative qualities that could be associated with green open spaces were perceived as significantly reducing their value to respondents (Figure 9). This suggests that open space planners and managers need to be aware of the potential for these to preclude the assignment of other positive green space values.

2 Mean Response

0 Noisy Unpleasant Unappealing Scary/Unsafe

Figure 9. Negative characteristics of green open spaces

3.2.4 Relationships between socio-demographics and general values for green open space

Table 3 displays correlations between a selection of socio-demographic variables and responses to Part 1 of the survey instrument (general values for green open space). Of particular interest are the positive correlations between age and (i) native plants and animals and (ii) health/therapeutic values. The negative correlation between the number of children a respondent has who are under 9 years of age and (i) native plants and animals value and (ii) nature appreciation activities could be because parents of young children do not have capacity to focus on biocentric environmental concerns. The more intuitive result however is the positive relationship between children <9 and the stated importance of children’s play activities. The number of children <9 is also negatively related to the importance of noise as a negative quality of open space. The final interesting result from this table is that income is negatively correlated with responses for all values/activities. This could be explained by full- time workers having less time to spend in green open spaces in their local area.

Table 3 Spearman’s Rank Correlation Coefficient (Spearman’s Rho) Years in Children Children Q Attribute Age LGA 10-17 <9 Education Income 1.1.1 Aesthetic/Scenic 0.130* -0.015 -0.051 -0.140** 0.120* -0.035 Activity/Physical 1.1.2 Exercise 0.010 -0.082 0.046 0.021 0.059 -0.089 Native Plants and 1.1.3 Animals 0.110* 0.027 0.035 -0.157** 0.001 -0.019 1.1.4 Nature 0.069 -0.051 0.040 -0.100 0.098 -0.021 Cultural 1.1.5 Significance 0.144** 0.003 -0.050 -0.133** 0.041 -0.046 1.1.6 Health/Therapeutic 0.226*** 0.070 -0.071 -0.168** 0.089 -0.087 1.1.7 Social Interaction 0.080 -0.009 0.016 -0.036 0.010 -0.136 1.2.1 Casual Recreation -0.033 -0.012 -0.036 0.020 0.176*** 0.007 1.2.2 Exercise for Fitness 0.059 -0.009 -0.023 -0.016 0.017 -0.067 1.2.3 Social Activities -0.064 0.034 0.031 0.071 -0.052 -0.163** 1.2.4 Children's Play -0.066 0.021 -0.048 0.224*** 0.042 -0.053 Nature 1.2.5 Appreciation 0.077 0.002 -0.009 -0.146*** 0.068 -0.068 1.3.1 Unappealing 0.063 0.097 -0.006 -0.072 -0.067 0.018 1.3.2 Scary/Unsafe -0.137** -0.045 0.041 0.154** 0.072 -0.049 1.3.3 Noisy 0.138** 0.013 -0.005 -0.229*** -0.048 0.037 1.3.4 Unpleasant 0.038 0.056 -0.047 -0.069 -0.004 -0.127** * = significant at <0.05, ** = significant at <0.01, *** = significant at <0.001 Q= Question number in survey book – refer to methods section for question, or Appendix A for full survey. Strength and direction of correlation are the most important.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 31

Table 4 below outlines relationships between categorical socio-demographic variables and responses to Part 1 of the survey (general green open space values). Differences in responses between groups were identified via Chi-square tests. Females gave higher responses for many green open space values and important activities than males. Members of conservation volunteer groups also had higher responses for nature and native plants and animal values, yet lower ratings for children’s play activities in green open spaces. Students were also found to respond less positively to children’s play activities. Children’s play was however most important to public housing tenants, with people holding a mortgage the second highest housing status group for this value. People who own their home outright are most likely to be concerned about green open spaces being noisy or unpleasant, while public housing tenants were least concerned about this.

Table 4 Relationships between socio-demographics and general values Principal Place of Volunteer Housing Dwelling Gender Residence Group Member Occupation Status Type Aesthetic/Scenic 5.78 3.118 5.968 38.131 18.353 7.656 Activity/Physical Exercise 8.419 3.139 5.165 19.643 5.838 6.925 Native Plants and Animals 14.112** 1.438 10.176* 26.813 11.788 7.182 Nature 11.4* 2.861 12.398** 28.356 5.529 5.53 Cultural Significance 18.512** 1.562 3.712 49.76 16.404 9.812 Health/Therapeutic 15.639** 3.183 9.178 62.784** 18.194 16.087 Social Interaction 12.417* 2.833 2.407 33.888 20.139 18.829 Casual Recreation 22.297*** 1.056 2.96 49.786 16.745 3.66 Exercise for Fitness 4.051 3.624 1.074 24.091 11.68 13.514 Social Activities 11.636* 5.904 0.73 28.912 14.581 17.996 Children's Play 7.507 1.357 11.459* 67.021** 30.322* 16.732 Nature Appreciation 7.052 0.744 11.275 46.076 13.95 13.468 Unappealing 6.954 6.898 9.837* 24.219 16.709 14.611 Scary/Unsafe 15.117** 7.634 7.235 35.727 11.56 11.598 Noisy 1.287 0.903 2.103 50.055 39.88** 10.108 Unpleasant 11.281* 5.086 14.103** 39.102 31.965* 14.697 Chi -sq test statistics are reported, along with significance levels. * = significant at <0.05, ** = significant at <0.01, *** = significant at <0.001

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 32

3.3 Community Satisfaction for Green Open Space

The dominant rating by survey respondents of their level of satisfaction with the amount (figure 10), quality (figure 11) and accessibility (figure 12) of green open space was ‘relatively satisfied’. This is encouraging for the municipalities in charge of the planning and maintenance of green space. However, there is scope to improve the proportion of respondents who are ‘totally satisfied’, for all four suburbs.

70 Charlestown Toronto

60 Nelson Bay Raymond Terrace 50 40 30 20 Percentage of Suburb Survey Respondents Survey of Suburb Percentage 10 0

Totally Satisfied Relatively Satisfied Relatively Unsatisfied Completely Unsatisfied

Neither Satisfied Nor Unsatisfied Figure 10. Satisfaction with amount of green open space, all respondents per suburb

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 33

60 Charlestown Toronto Nelson Bay

50 Raymond Terrace 40 30 20 Percentage of Suburb Survey Respondents Survey of Suburb Percentage 10 0

Totally Satisfied Relatively Satisfied Relatively Unsatisfied Completely Unsatisfied

Neither Satisfied Nor Unsatisfied Figure 11. Satisfaction with the Quality of Green Open Space

60 Charlestown Toronto Nelson Bay 50 Raymond Terrace 40 30 20 Percentage of Suburb Survey Respondents Survey of Suburb Percentage 10 0

Totally Satisfied Relatively Satisfied Relatively Unsatisfied Completely Unsatisfied

Neither Satisfied Nor Unsatisfied Figure 12. Accessibility of green open space

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 34

3.4 General Green Open Space Values

As a way of the structure of green open space values and reducing the number of variables in the dataset, exploratory factor analysis was employed on the Part 1 survey responses (general green open space benefits). A four-factor solution was identified as most appropriate, and the factor loadings of each of the items are outlined in Table 5 below. Items with factor loadings >0.4 are highlighted in bold according to common practice. Since only one loading value >0.4 was identified for each factor, these can be considered the dominant items contributing to each factor.

The exploratory factor analysis identified 4 distinct factors: • F1 Nature/culture; • F2 Negative; • F3 Social; and • F4 Activity.

Table 5 Factor loadings for general open space benefits items, part 1 of the survey F1: F2: F3: F4: Nature/Culture Negative Social Activity Values Aesthetic/Scenic Value 0.438 0.153 0.183 0.116 Activity/Physical Exercise 0.216 0.058 0.290 0.552 Native Biota 0.728 0.004 -0.025 0.076 Nature 0.815 0.040 0.065 0.118 Cultural Significance 0.624 0.011 0.343 0.075 Health/Therapeutic 0.524 0.065 0.321 0.186 Social Interaction 0.241 0.038 0.639 0.137 Important Casual Recreation 0.238 0.141 0.315 0.429 Activities Exercise for fitness 0.095 0.011 0.169 0.978 Social activities 0.152 0.049 0.718 0.242 Children’s play 0.136 0.036 0.647 0.143 Nature appreciation 0.660 -0.043 0.227 0.086 Negative Unappealing 0.026 0.727 0.010 0.033 Qualities Scary/Unsafe -0.051 0.639 0.083 0.026 Noisy 0.203 0.641 -0.059 0.035 Unpleasant -0.009 0.693 0.102 0.050

The inclusion of ‘value’ and ‘important activity’ items on the same factors suggests that these two categories are similar psychological constructs and perceived similarly by respondents. The key differences are about the type of values and activities. Nature and culture items tended to load on the same factor, suggesting that people who care about the environmental benefits of

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 35

open spaces also value cultural, aesthetic and therapeutic benefits. Negative qualities were found to load strongly on their own factor. This implies that people who care strongly about one negative feature are also likely to be concerned about others. This is most likely related to personality orientations of individuals.

3.5 Mapping Values and Uses of Green Open Spaces

The following results detail the findings from Parts 2 and 3 of the survey: those components that involved mapping values, activities and negative qualities across the landscape.

3.5.1 Abundance of Dots at Different Scales

Figure 13 shows that maps at the suburb scale (Map 2) received significantly more value dots than those at the LGA scale (Map 1) for all value attributes (Chi-squared = 404.4, df = 15, p-value < 0.001). This difference was especially pronounced for values such as activity/physical exercise and social interaction, and activities such as casual recreation and exercise for fitness. Nature value had a similar number of value dots assigned to maps at both scales, suggesting that people’s appreciation for nature in green open spaces is not restricted to their immediate local area. It is interesting that there is greater variability among values and activities for the mapping exercise than there is for Part 1 of the survey, which looked at these attributes in a general sense. For both suburb and LGA scale maps, native plants and animals value, nature value and nature appreciation activities are comparatively lower than in Part 1 responses, suggesting that there is an opportunity for greater provision of areas that maximise these values.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 36

Unpleasant

LGA Suburb

Noisy

Scary/Unsafe

Unappealing

Appreciation Nature

Play Childrens

Activities Social

Fitness For Exercise

Recreation Casual

Interaction Social

Dot Frequency LGA v Suburb Health/Therapeutic

Significance Cultural

Nature

Animals and Plants Native

Exercise Activity/Physical

Aesthetic/Scenic

800 600 400 200 1000 Dots of Number

Figure 13. Dot abundance for each attribute; LGA and suburb scale

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 37

The factor scores as per (table 5 and section above) were related to the abundance of mapped dots at the suburb and LGA scales. Many of the mapped attributes at the suburb scale were found to be significantly correlated with participants’ responses to the general open space survey questions in Part 1 of the questionnaire.

These relationships are outlined in Tables 6 and 7 on the following page.

Many of the relationships observed are as expected; the abundance of dots is related to the strength of similar general values respondents have for green open space. It is interesting however that the abundance of dots for native plants and animals value, nature value and nature appreciation activities was related not only to the nature/culture factor but also the social factor at the suburb scale. This emphasises a potential compatibility between conservation values and social values.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 38

Table 6 Local Government Scale Map (Map 1) – Pearson’s R correlation coefficient Factors of General Values (Not Mapped) Nature/Culture Negative Social Activity Aesthetic 0.090 -0.068 -0.032 0.016 Activity/Exercise 0.132* 0.035 -0.049 0.105* Native Biota 0.173*** -0.049 0.006 -0.029

Nature 0.155** -0.034 0.046 -0.010 Cultural Significance 0.235*** -0.012 0.141** -0.104* Health/Therapeutic 0.137** 0.008 0.004 0.009 Social Interaction 0.080 0.008 0.144** 0.065 Casual Recreation 0.008 -0.009 -0.014 0.074 Exercise for Fitness 0.001 -0.008 -0.005 0.127* Social Activities 0.038 -0.026 0.047 0.028 Children's Play 0.007 -0.007 0.078 0.014 Nature Appreciation 0.171** -0.126* -0.036 -0.013

Mapped Values (number of dots) Unappealing 0.070 0.026 0.025 -0.071 Scary/Unsafe 0.066 -0.015 0.044 -0.050 Noisy 0.051 0.035 0.033 -0.052 Unpleasant 0.052 0.067 -0.015 0.000 * = significant at <0.05, ** = significant at <0.01, *** = significant at <0.001 n = 371

Table 7 Suburb Scale Map (Map 2) – Pearson’s R correlation coefficient – Pearson’s R Factors of General Values (Not Mapped) Nature/Culture Negative Social Activity Aesthetic 0.207*** 0.087 0.043 -0.035 Activity/Exercise 0.121* 0.063 0.069 0.057 Native Biota 0.276*** 0.063 0.115* 0.105*

Nature 0.239*** -0.009 0.131* 0.118* Cultural Significance 0.171*** 0.008 0.149** 0.017 Health/Therapeutic 0.221*** 0.047 0.118* 0.059 Social Interaction 0.105* -0.028 0.224*** 0.078 Casual Recreation 0.180*** 0.031 0.097 0.143** Exercise for Fitness 0.089 -0.002 0.115* 0.194*** Social Activities 0.089 0.037 0.177*** 0.080 Children's Play 0.023 -0.023 0.267*** 0.060 Nature Appreciation 0.251*** 0.035 0.124* 0.058

Mapped Values (number of dots) Unappealing -0.060 -0.014 0.053 -0.094 Scary/Unsafe -0.112* 0.082 0.045 -0.059 Noisy 0.053 0.018 -0.043 -0.139** Unpleasant 0.015 0.079 0.021 -0.014 * = significant at <0.05, ** = significant at <0.01, *** = significant at <0.001 n = 371

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 39

3.5.2 PPGIS – Mapping Values for Green Open Spaces

Figures 14 to 17 presented on the following pages show the dot abundance (value density) for all attributes for each suburb (Map 2 of the survey), with densities aggregated across a 100m grid. Figures 18-21 show the dot abundance per 100m2 for Nature, Native Biota and Nature activities for each suburb. Given the large number of value heat maps produced, we have chosen to present those for only a selection of value attributes. Maps showing all individual attributes for each suburb are available upon request from the authors.

It is clear from the dot abundance figures that the highest densities of all green space value attributes are located near water bodies. It is also evident that the ‘nature’ themed attributes do not necessarily align with those areas that might be considered of highest ecological value (e.g. large, well-connected, undisturbed parks). It is likely that ‘nature’ values are also related to reasonable levels of access and other positive features (such as aesthetic values).

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 40

Figure 14. Dot abundance for all attributes, Nelson Bay

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 41

Figure 15. Dot abundance for all attributes, Charlestown

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 42

Figure 16. Dot abundance for all attributes, Toronto

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 43

Figure 17. Dot abundance for all attributes, Raymond Terrace

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 44

Nature values Native plants and animals values

Nature appreciation activities

Figure 18. Dot abundance per 100m2 for nature and native biota values, and nature activities, Nelson Bay FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 45

Nature values Native plants and animals values

Nature appreciation activities

Figure 19. Dot abundance per 100m2 for nature and native biota values, and nature activities, Toronto

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 46

Native plants and animals values Nature values

Nature appreciation activities

Figure 20. Dot abundance per 100m2 for nature and native biota values, and nature activities, Raymond Terrace

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 47

Nature values Native plants and animals values

Nature appreciation activities

Figure 21. Dot abundance per 100m2 for nature and native biota values, and nature activities, Charlestown

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 48

3.6 Understanding People’s Favourite Green Open Spaces

3.6.1 Factors Relating to People’s Favourite Green Open Spaces

Figures 22 and 23 outline the importance of different characteristics that contribute to a green space being identified as a respondent’s favourite at the LGA and suburb scale. It is evident that there is substantial variability for each factor (as shown by the standard deviation error bars), highlighting the fact that people have favourite places for different reasons. The presence of a water body and beautiful views consistently rated highly amongst the possible attributes of the green space. It is also interesting that ‘quiet’ is rated lowest for both the LGA and suburb scale favourite places. Of particular importance to biodiversity conservation is the fact that the presence of birds and other wildlife and the environmental importance of a place was overall rated more highly than the maintenance and facilities of a site – those factors that generally concern municipal open space managers. Finally, the importance of these environmental features are just as high at the suburb scale as they are at the LGA scale, even though the large national parks were more identifiable on the LGA scale map.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 49

Factors Related to Favourite Places in Suburb

3 Survey Response Survey 2

0 Quiet Clean Close by Water body Water Mature trees Built features Good facilities Beautiful views − maintained Well Easily accessible Pleasant memories Lots of green areas Enjoyable for children for Enjoyable Birds and other wildlife Environmental importanceEnvironmental Figure 22. Factors related to favourite places in LGA

Factors Related to Favourite Places in LGA

3 Survey Response Survey 2

Figure 23. Factors Related to Favourite Places in Suburb

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 50

3.7 Factors Influencing Mapped Green Open Space Values (Suburb)

3.7.1 Environmental and Landscape Factors

This section outlines the effect of a range of landscape and environmental factors on the abundance of value dots assigned to individual parks. As discussed in the methods, the abundance of mapped dots in each park was calculated by intersecting the park polygons with the digitised dots. This section concentrates on park polygons at the suburb scale (Map 2) since this was the scale where particular parks could be identified and their attributes calculated. Conversely, the LGA scale map was too coarse to provide meaningful information about the specific relationships between environmental variables and mapped values.

3.7.1.1 Park Area

The relationship between the area of a park and the number of mapped dots was analysed via zero-inflated Poisson modelling. The formula is outlined below:

Y ~ P(λ) logλ = a + bX + cX2 where Y = number of dots in a park for a given value attribute P = Poisson distribution λ = the expected count (Poisson mean) a = model intercept X = area X2 = quadratic area term. N.B. Since X was standardised for the analysis, calculating the true X from the formula will require reverse transformation.

Table 8 outlines the results of the modelling, with the Pseudo R2 value (McFadden’s Pseudo R2) providing a measure of the goodness of fit. Each model for each value attribute is also displayed graphically in the following figures.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 51

Table 8 Parameters for Zero-inflated Poisson models of park value dot abundance and park area

Attribute b c Pseudo R2 λ Mean CI (upper) CI (lower) λ Mean CI (upper) CI (lower) 01 Aesthetics 0.88 0.76 1.00 -0.15 -0.18 -0.13 0.10 02 Activity/Physical Exercise Value 0.62 0.54 0.70 -0.06 -0.07 -0.05 0.20 03 Native Plants and Animals Value 0.88 0.71 1.06 -0.13 -0.17 -0.10 0.11 04 Nature Value 1.00 0.78 1.22 -0.14 -0.18 -0.10 0.13 05 Cultural Significance Value 0.76 0.45 1.08 -0.19 -0.28 -0.10 0.04 06 Health/Therapeutic Value 1.07 0.87 1.27 -0.17 -0.21 -0.13 0.13 07 Social Interaction Value 1.02 0.88 1.15 -0.17 -0.21 -0.14 0.15 08 Casual Recreation Activities 1.09 0.96 1.23 -0.20 -0.23 -0.17 0.16 09 Exercise for Fitness 1.18 1.03 1.33 -0.19 -0.22 -0.16 0.20 10 Social Activities 0.54 0.37 0.71 -0.10 -0.16 -0.05 0.04 11 Children's Play 0.43 0.30 0.56 -0.09 -0.13 -0.06 0.03 12 Nature Appreciation Activities 0.94 0.72 1.15 -0.14 -0.18 -0.10 0.11 13 Unappealing 0.76 0.55 0.96 -0.11 -0.16 -0.07 0.07 14 Scary/Unsafe 0.75 0.44 1.06 -0.12 -0.18 -0.05 0.04 15 Noisy 0.76 0.37 1.15 -0.14 -0.23 -0.05 0.04 16 Unpleasant 1.58 1.22 1.95 -0.28 -0.37 -0.18 0.18 Note: All models have been developed for parks <40ha in size due to insufficient data for larger values

Below are the plots for all points compared to area for each individual attribute (Figure 24). Each value is positively related to green space area, with the goodness of fit of the model strongest for activity/physical exercise value and exercise for fitness activities. The quadratic term was significant for all models, suggesting a non-linear effect of area (strongest effects are for smaller parks).

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 52

01 Aesthetic Value 02 Activity/Physical Exercise Value 140 160 110 120 80 90 60 50 Number of Dots Number of Dots 30 20 0 0

0 1 5 10 20 40 0 1 5 10 20 40 Park Area (ha) Park Area (ha)

03 Native Biota Value 04 Nature Value 50 50 40 40 30 30 20 20 Number of Dots Number of Dots 10 10 0 0

0 1 5 10 20 40 0 1 5 10 20 40 Park Area (ha) Park Area (ha)

05 Cultural Value 06 Health Value 80 30 60 20 40 10 Number of Dots Number of Dots 20 0 0

0 1 5 10 20 40 0 1 5 10 20 40 Park Area (ha) Park Area (ha)

07 Social Interaction Value 08 Casual Recreation 110 80 80 60 60 40 40 Number of Dots Number of Dots 20 20 0 0

0 1 5 10 20 40 0 1 5 10 20 40 Park Area (ha) Park Area (ha)

Figure 24 Park value dot abundance and park area per attribute The solid line indicates the fitted model, with the dashed lines denoting the upper and lower confidence intervals.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 53

09 Exercise for Fitness 10 Social Activities 100 100 80 80 60 60 40 40 Number of Dots Number of Dots 20 20 0 0

0 1 5 10 20 40 0 1 5 10 20 40 Park Area (ha) Park Area (ha)

11 Children's Play 12 Nature Appreciation 80 40 60 30 40 20 Number of Dots Number of Dots 20 10 0 0

0 1 5 10 20 40 0 1 5 10 20 40 Park Area (ha) Park Area (ha)

13 Unappealing 14 Scary/Unsafe 20 10 5 10 Number of Dots Number of Dots 0 0

0 1 5 10 20 40 0 1 5 10 20 40 Park Area (ha) Park Area (ha)

15 Noisy 16 Unpleasant 15 10 10 5 5 Number of Dots Number of Dots 0 0

0 1 5 10 20 40 0 1 5 10 20 40 Park Area (ha) Park Area (ha)

Figure 24. (continued) Park value dot abundance and park area per attribute The solid line indicates the fitted model, with the dashed lines denoting the upper and lower confidence intervals.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 54

3.7.1.2 Percentage Vegetation in Park

Table 9 shows the effect of percentage of vegetation cover in park on number of dots per attribute. The vegetation value was calculated by intersecting park polygons with the vegetation cover dataset obtained from ERIN at a resolution of 20m. As for park area, the Poisson model was as follows: Y ~ P(λ) logλ = a + bX + cX2 where Y = number of dots in a park for a given value attribute P = Poisson distribution λ = the expected count (Poisson mean) a = model intercept X = per cent vegetation in park X2 = quadratic vegetation term.

Table 9 Parameters for Zero-inflated Poisson models of park value dot abundance and percentage of vegetation in the park.

Attribute b c Pseudo R2 Mean CI (lower) CI (upper) Mean CI (lower) CI (upper) 01 Aesthetics 0.20 0.11 0.29 -0.20 -0.30 -0.10 0.01 02 Activity/Physical Exercise Value 0.08 0.00 0.16 -0.30 -0.39 -0.21 0.02 03 Native Plants and Animals Value 0.71 0.52 0.90 -0.40 -0.54 -0.26 0.05 04 Nature Value 0.63 0.40 0.86 -0.39 -0.56 -0.22 0.03 05 Cultural Significance Value 0.56 0.33 0.78 -0.58 -0.83 -0.34 0.07 06 Health/Therapeutic Value 0.36 0.20 0.51 -0.32 -0.47 -0.17 0.02 07 Social Interaction Value -0.07 -0.18 0.04 -0.35 -0.48 -0.22 0.02 08 Casual Recreation Activities 0.11 0.01 0.22 -0.42 -0.54 -0.31 0.03 09 Exercise for Fitness 0.27 0.16 0.38 -0.27 -0.38 -0.15 0.02 10 Social Activities 0.20 0.06 0.33 -0.78 -0.93 -0.62 0.10 11 Children's Play -0.07 -0.19 0.05 -0.54 -0.69 -0.40 0.05 12 Nature Appreciation Activities 0.47 0.26 0.68 -0.29 -0.45 -0.12 0.02 13 Unappealing 0.31 0.15 0.47 -0.20 -0.35 -0.05 0.02 14 Scary/Unsafe 0.40 0.15 0.65 -0.33 -0.58 -0.08 0.02 15 Noisy -0.22 -0.53 0.08 -0.17 -0.49 0.14 0.01 16 Unpleasant -0.43 -0.70 -0.17 0.11 -0.19 0.40 0.02

Figure 25 shows these relationships graphically. Note that many of these plots have log scale y- axes for ease of interpretation. Although most variables had a unimodal relationship (a peak response corresponding with a mid-level of vegetation cover), native biota value, nature value and nature appreciation activities were all strongly positively related to park vegetation cover, with the highest values being associated with higher vegetation levels. Conversely, social interaction value and many recreation activities peak at mid-levels of vegetation cover. This is likely explained by parks with very high levels of vegetation cover offering insufficient clear space for these activities, while parks with low levels of vegetation cover may be less appealing.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 55

01 Aesthetic Value 02 Activity/Physical Exercise Value 100 100 50 50 10 10 5 5 Number of Dots Number of Dots 0 0

0 20 40 60 80 100 0 20 40 60 80 100 Park Vegetation (%) Park Vegetation (%)

03 Native Biota Value 04 Nature Value 50 50 10 10 5 5 Number of Dots Number of Dots 0 0

0 20 40 60 80 100 0 20 40 60 80 100 Park Vegetation (%) Park Vegetation (%)

05 Cultural Value 06 Health/Therapeutic Value 50 100 50 10 10 5 5 Number of Dots Number of Dots 0 0

0 20 40 60 80 100 0 20 40 60 80 100 Park Vegetation (%) Park Vegetation (%)

07 Social Interaction Value 08 Casual Recreation 100 100 50 50 10 10 5 5 Number of Dots Number of Dots 0 0

0 20 40 60 80 100 0 20 40 60 80 100 Park Vegetation (%) Park Vegetation (%)

Figure 25. Park value dot abundance and percentage vegetation cover per attribute

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 56

09 Exercise for Fitness Social Activities 100 100 50 50 10 10 5 5 Number of Dots Number of Dots 0 0

0 20 40 60 80 100 0 20 40 60 80 100 Park Vegetation (%) Park Vegetation (%)

11 Children's Play 12 Nature Appreciation 50 100 50 10 10 5 5 Number of Dots Number of Dots 0 0

0 20 40 60 80 100 0 20 40 60 80 100 Park Vegetation (%) Park Vegetation (%)

13 Unappealing 14 Scary/Unsafe 10 15 8 10 6 4 5 Number of Dots Number of Dots 2 0 0

0 20 40 60 80 100 0 20 40 60 80 100 Park Vegetation (%) Park Vegetation (%)

15 Noisy 16 Unpleasant 12 12 10 10 8 8 6 6 4 4 Number of Dots Number of Dots 2 2 0 0

0 20 40 60 80 100 0 20 40 60 80 100 Park Vegetation (%) Park Vegetation (%)

Figure 25. (continued) Park value dot abundance and percentage vegetation cover per attribute

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 57

3.7.1.3 Park management category

The influence of green open space management category on mapped values was also analysed via zero-inflated Poisson modelling. The three management categories (simplified from data provided by the two municipalities) were (i) general parks, (ii) natural areas, and (iii) sportsfields. While this typology is simpler than that used in practice, it enabled data from all suburbs to be compiled reasonably.

As park management type was a categorical variable, the model formula was as follows:

Y ~ P(λ)

logλ = bi[category]

where Y = number of dots in a park for a given value attribute P = Poisson distribution λ = the expected count (Poisson mean) i = 1…3 (park management categories) b = effect size of park management category.

The results of the models are presented in Table 10

Table 10 Parameters for Zero-inflated Poisson models of park value dot abundance and park management category General Natural Sportsfield Attribute λ Mean CI (upper) CI (lower) λ Mean CI (upper) CI (lower) λ Mean CI (upper) CI (lower) 01 Aesthetics 13.72 14.99 12.56 5.17 5.64 4.73 3.60 3.92 3.29 02 Activity/Physical Exercise Value 10.67 11.70 9.72 5.18 5.68 4.72 10.00 10.97 9.12 03 Native Plants and Animals Value 6.28 7.34 5.38 4.77 5.57 4.09 1.95 2.27 1.67 04 Nature Value 6.53 7.64 5.58 4.29 5.02 3.66 1.08 1.27 0.92 05 Cultural Significance Value 3.32 4.13 2.67 4.64 5.77 3.72 1.51 1.88 1.21 06 Health/Therapeutic Value 6.87 7.83 6.03 2.71 3.09 2.38 2.38 2.71 2.09 07 Social Interaction Value 9.22 10.22 8.33 2.49 2.76 2.25 5.82 6.44 5.25 08 Casual Recreation Activities 9.89 10.86 9.00 3.83 4.21 3.49 4.99 5.48 4.54 09 Exercise for Fitness 7.40 8.33 6.57 5.48 6.17 4.87 7.04 7.92 6.25 10 Social Activities 14.00 15.43 12.70 2.90 3.20 2.63 1.78 1.97 1.62 11 Children's Play 11.18 12.28 10.18 2.14 2.35 1.95 3.73 4.10 3.40 12 Nature Appreciation Activities 5.10 6.03 4.32 4.14 4.89 3.50 1.61 1.90 1.36 13 Unappealing 2.77 3.32 2.32 2.58 3.09 2.16 2.26 2.71 1.89 14 Scary/Unsafe 1.99 2.64 1.50 1.84 2.44 1.39 1.06 1.41 0.80 15 Noisy 1.76 2.42 1.28 1.03 1.42 0.75 2.41 3.31 1.75 16 Unpleasant 1.23 1.84 0.81 0.69 1.03 0.46 2.81 4.22 1.87

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 58

The following plots in Figure 26 display the 16 attributes against park management category per attribute. The expected counts for each category are shown as black circles, with the upper and lower error bounds. The real counts are overlaid in grey. Although there is substantial variability within each category, we note that ‘natural’ areas are appreciated for values other than strictly environmental concerns. For example, they contain only a marginally lower count mean for ‘exercise for fitness activities’ than both general parks and sportsfields. In addition, we find that general parks are modelled as containing as many or more mapped dots for native biota value, nature value and nature appreciation activities than parks classed as natural areas. This highlights the fact that people appreciate nature and biodiversity in a range of park management categories or types and do not restrict these values to areas that are managed primarily for environmental outcomes.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 59

01 Aesthetic Value Activity/Physical Exercise Value 100 100 50 50 10 10 5 5 Number of Dots in Park Number of Dots in Park 0 0

General Natural Sportsfield General Natural Sportsfield Park Category Park Category

03 Native Plants and Animals 04 Nature Value 50 50 10 10 5 5 Number of Dots in Park Number of Dots in Park 0 0

General Natural Sportsfield General Natural Sportsfield Park Category Park Category

05 Cultural Significance Value 06 Health/Therapeutic Value 50 100 50 10 10 5 5 Number of Dots in Park Number of Dots in Park 0 0

General Natural Sportsfield General Natural Sportsfield Park Category Park Category

07 Social Interaction Value 08 Casual Recreation 100 100 50 50 10 10 5 5 Number of Dots in Park Number of Dots in Park 0 0

General Natural Sportsfield General Natural Sportsfield Park Category Park Category

Figure 26. Park value dot abundance and park management category per attribute

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 60

09 Exercise for Fitness 10 Social Activities 100 100 50 50 10 10 5 5 Number of Dots in Park Number of Dots in Park 0 0

General Natural Sportsfield General Natural Sportsfield Park Category Park Category

11 Children's Play 12 Nature Appreciation 50 100 50 10 10 5 5 Number of Dots in Park Number of Dots in Park 0 0

General Natural Sportsfield General Natural Sportsfield Park Category Park Category

13 Unappealing 14 Scary/Unsafe 10 15 8 10 6 4 5 Number of Dots in Park Number of Dots in Park 2 0 0

General Natural Sportsfield General Natural Sportsfield Park Category Park Category

15 Noisy 16 Unpleasant 10 12 10 8 8 6 6 4 4 Number of Dots in Park Number of Dots in Park 2 2 0 0

General Natural Sportsfield General Natural Sportsfield Park Category Park Category

Figure 26. (continued) Park value dot abundance and park management category per attribute

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 61

3.7.1.4 Park distance from water

The influence of the distance of a park from water on the assignment of value dots was assessed via zero-inflated Poisson modelling. This dataset considered large water bodies but not smaller streams and creeks, with the results shown in Table 11. The formula used was:

Y ~ P(λ) logλ = a + bX + cX2 where Y = number of dots in a park for a given value attribute P = Poisson distribution λ = the expected count (Poisson mean) a = model intercept X = = Distance of park from water (m) X2 = quadratic distance form water term.

Table 11 Parameters for Zero-inflated Poisson models of park value dot abundance and park distance from water Attribute b c Pseudo,R2 Mean CI'(lower) CI'(upper) Mean CI'(lower) CI'(upper) 01#Aesthetics +1.21 +1.44 +0.97 0.40 0.18 0.63 0.23 02#Activity/Physical#Exercise#Value +0.61 +0.74 +0.49 0.53 0.38 0.69 0.13 03#Native#Plants#and#Animals#Value +0.43 +0.58 +0.28 +0.05 +0.23 0.13 0.07 04#Nature#Value +0.39 +0.58 +0.21 0.06 +0.15 0.27 0.04 05#Cultural#Significance#Value +0.84 +1.18 +0.50 +0.35 +0.74 0.04 0.08 06#Health/Therapeutic#Value +1.18 +1.44 +0.91 0.37 0.10 0.65 0.17 07#Social#Interaction#Value +1.00 +1.21 +0.79 0.24 0.01 0.47 0.21 08#Casual#Recreation#Activities +0.79 +0.95 +0.64 0.37 0.19 0.54 0.15 09#Exercise#for#Fitness +0.49 +0.63 +0.36 0.15 +0.02 0.32 0.07 10#Social#Activities +2.02 +2.49 +1.55 0.16 +0.21 0.53 0.25 11#Children's#Play +1.00 +1.27 +0.74 0.36 0.10 0.62 0.20 12#Nature#Appreciation#Activities +0.20 +0.39 +0.02 +0.17 +0.42 0.07 0.02 13#Unappealing 0.23 0.02 0.44 +0.17 +0.44 0.09 0.01 14#Scary/Unsafe +0.21 +0.51 0.09 +0.15 +0.51 0.22 0.02 15#Noisy +1.00 +1.71 +0.30 +0.49 +1.19 0.21 0.07 16#Unpleasant +0.15 +0.50 0.20 +0.44 +0.87 +0.02 0.02 N.B. Distance from water was only for parks <500m due to the minimal effect of the variable for parks further than this distance.

The following figure 27 shows the 16 attributes against park distance from water per attribute. For all positive values and activities, there was a negative relationship between distance from water and the number of dots in a park (i.e. parks closer to water were valued more highly). This was especially pronounced for distances <100m, with the steepest slopes associated with aesthetic values and social activities. Nature and native biota values were not as strongly related to distance from water.

FINAL REPORT: PLANNING FOR GREEN OPEN SPACE IN URBANISING LANDSCAPES 62

01 Aesthetic Value 02 Activity/Physical Exercise Value 100 100 50 50 10 10 5 5 Number of Dots Number of Dots 0 0