Idea Transcript
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
COPYRIGHT PAGE © 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 All photographs copyright © and were taken by Cathy Oke or Chris Ives unless otherwise indicated. 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 (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: 𝐷𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑓𝑟𝑜𝑚 𝐻𝑜𝑚𝑒 𝑡𝑜 𝑃𝑎𝑟𝑘 𝐶𝑒𝑛𝑡! = ( 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 0.005 0.000
Proportion of Respondents
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
20
40
60
House Townhouse Flat Other
0
Number of Survey Respondents
80
other forms of dwelling types.
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 0
20
40
60
80
No Yes
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 10 5 0
Number of Survey Respondents
Negative Nil $0−$10,399 $10,400−$15,599 $15,600−$20,799 $20,800−$31,199 $31,200−$41,599 $41,600−$51,999 $52,000−$64,999 $65,000−$77,999 $78,000−$103,999 $104,000+
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.
50 40 30 20 10 0
Number of Survey Respondents
60
Own Outright Own Mortgage Rent Public Housing Other
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’.
5
Mean Response
4 3 2 1
Social Interaction
Health/Therapeutic
Cultural Significance
Nature
Native Plants and Animals
Activity/Physical Exercise
Aesthetic/Scenic
0
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).
5
Mean Response
4
3
2
1
Nature Appreciation
Childrens Play
Social Activities
Exercise For Fitness
Casual Recreation
0
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.
5
Mean Response
4
3
2
1
Figure 9. Negative characteristics of green open spaces
Unpleasant
Noisy
Scary/Unsafe
Unappealing
0
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