-
We know little about how to design new, or improve or promote existing, urban green space for health and social outcomes.
-
Interventions should employ a dual approach that incorporates promotion and marketing of urban green space as well as changing the physical environment.
-
There is evidence to support a range of environmental, health and social benefits.
-
Little is known about the equity impact of urban green space interventions.
17.1 Introduction
17.2 Urban Green Space Interventions
17.2.1 What Are Urban Green Space Interventions?
17.2.2 Park-Based Interventions
Reference | Study design | Population | Intervention | Outcome |
---|---|---|---|---|
Cohen et al. (2009a) Los Angeles, CA, USA | Quasi-experiment: controlled, pre-post design | Predominantly Latino and African-American and low-income neighbourhood | 5 parks (mean 8 acres) underwent major improvements including new/improved gyms, picnic areas, walking paths, playgrounds, watering and landscaping (cost: >$1 m each) Parks ranged from 3.4 to 16 acres (mean 8 acres) and served an average of 67,000 people within a 1-mile radius. Parks contained multi-purpose fields; playgrounds; gymnastics areas; and picnic and lawn areas. |
−ve: Overall park use and PA declined in both intervention and control parks |
Cohen et al. (2009b) Los Angeles, CA, USA | Quasi-experiment: controlled, pre-post design | Youths and seniors living within 2 mile radius of parks | 2 parks (48–67 acres) underwent renovations: (1) improvements to skate park surfaces only (cost $3.5 m) (2) improvements to entrance, courtyard areas and gymnasium of senior centre (cost $3.3 m) |
−ve: 510% increase in skate park use compared to 77% in comparison skate park Substantially fewer users of senior centre |
Quigg et al. (2011) Dunedin, New Zealand | Quasi-experiment: controlled, pre-post design | Children aged 5–10 years from the local community | 2 community playgrounds: (1) playground had 10 new components installed, including play equipment, seating, additional safety surfacing, and waste facilities; (2) playground had 2 new play equipment pieces installed |
−ve: No statistically sig. difference in total daily PA compared with control. |
Cohen et al. (2012) Los Angeles, CA, USA | Quasi-experiment: controlled, pre-post design | Residents within 1 mile radius of intervention parks | 12 parks (mean 14 acres) involving installation of Family Fitness zones (outdoor gyms), 8 pieces of equipment at each park (average cost $45,000 for each park), mean park size 14.4 acres (range, 1–29 acres); served an average of 40,964 individuals within 1-mile radius |
−ve: Park usage increased by 11% compared to control parks (not sig.) |
Veitch et al. (2012) Victoria, Australia | Quasi-experiment: controlled, pre-post design | Most disadvantaged decile in state of Victoria | 1 park (size 25,200 m2): involving establishment of a fenced leash-free area for dogs (12,800 m2); an all-abilities playground; a 365 m walking track; BBQ area; landscaping; fencing to prevent motor vehicle access to the park | +ve: Sig. increase from pre to post-improvement in number of park users for intervention park (T1 = 235, T3 = 985) and number of people walking (T1 = 155, T3 = 369) and being vigorously active (T1 = 38, T3 = 257) |
Bohn-Goldhaum et al. (2013) Sydney, Australia | Quasi-experiment: controlled, pre-post design | 2–12 year olds and their parents or care givers; low socioeconomic neighbourhood | 1 park underwent renovations: new children’s play equipment, upgrading paths, adding new greenery, lighting and facilities (e.g., park furniture), green space was created by opening the adjacent sports field to public use |
−ve: No sig. difference between parks for usage or the number of children engaging in MVPA at follow-up. In the intervention park the number of girls engaging in MVPA significantly decreased (p = 0.04) between baseline and follow-up |
Cohen et al. (2014) Los Angeles, CA, USA | Quasi-experiment: post data only | Residents living within 0.5 mile radius of parks; minority populations | Creation of 3 pocket parks (0.15–0.32 acres) from vacant lots and undesirable urban parcels; playground equipment and benches installed, walking path developed around the perimeter, all fenced and enclosed by lockable gates (average cost $1 m per park funded by local non-profit groups) |
−ve: Pocket parks were used as frequently or more often than playground areas in neighborhood parks. However, they were vacant during the majority of observations |
Peschardt and Stigsdotter (2014) Copenhagen, Denmark | Natural experiment: pre-post design | 52% male; 88% Danish | A pocket park (932 m2) in a dense urban area was redesigned to increase seating areas and walking trails |
−ve: No sig. change in number of park users but demographics of park users changed slightly with more men, people aged 15–29 and more educated people using the park |
Quasi-experiment: controlled, pre-post design | Adolescents (12–15 years) and adults in severely deprived neighbourhoods | Dutch District Approach (five million euros): new public parks replacing vacant land (n = 9), refurbishing existing parks (n = 9), n = 6 improving paths, drainage, landscaping, planting flower bulbs in front yards; constructing wall gardens; greening streets, developing a greenway |
−ve: Intervention areas did not show more favourable changes in PA and general health compared to all the different groups of control areas for adults | |
NSW Health (2002) Sydney, Australia | Quasi-experiment: controlled, pre-post design | Residents aged 25–65 years | 3 types of interventions in 3 parks: promoting PA and park use (via advertisements, walking maps), park modifications (signage, greening, improved paths, new playground) and the establishment of walking groups | +ve: Intervention group more likely to have walked in the 2 weeks prior to follow-up than control. Sig. group by gender interaction indicated, intervention males were 2.8 times more likely to walk than were males in the control ward |
Tester and Baker (2009) San Francisco, CA, USA | Quasi-experiment: controlled, pre-post design | Resource poor neighbourhoods; primarily Latino community | Major renovations to two parks: lighting, fencing, artificial turf, landscaping, picnic benches, goal posts, walkways | +ve: Sig. increases of greater than fourfold magnitude among children and adults of both genders at the intervention park playfields, but not in the control park; sig. park use in non-play fields |
Cohen et al. (2013) Los Angeles, CA, USA | RCT: parks randomized to 3 study arms (17 parks per study arm) | Parks users and residents living within 1 mile radius of park | 2 intervention groups: (1) Park Director only; (2) Park Advisory Board-Park Director Involved in all aspects of research and in using baseline results to design park-specific interventions to increase park use and PA; Park Directors received five training sessions from a marketing consultant Each park received $4000 to spend on signage; promotional incentives; outreach and support for group activities | +ve: In both intervention parks, PA increased, generating an estimated average of 600 more visits/week/park, and 1830 more MET-hours of PA/week/park |
Ward Thompson et al. (2013) Glasgow, Scotland, United Kingdom | Quasi-experiment: controlled, pre-post design | High socioeconomic deprivation areas with woods/green space within 500 m of the community | Regeneration of local community: construction of improved footpaths; clearing rubbish and signs of vandalism; signage and entrance gateways; silvicultural work to improve appearance and safety of trees and vegetation (improve views and visibility); publicity and group activities to encourage knowledge of woodlands and opportunities for use. | +ve: Quality of life sig. increased in both neighbourhoods (more in intervention) over time and a sig. difference in quality of the physical environment between sites in 2006 but not 2009. Sig. differences in perceptions of safety (p < 0.05) in the intervention site over time, compared with no sig. change in the control |
King et al. (2015) Denver, CO, USA | Quasi-experiment: pre-post design | Residents of transitional housing (homeless and refugees) | Transformation of 2-acres of undeveloped green space into a recreational park and community garden The new park had clearly defined recreational spaces including a multi-purpose playing field; playground equipment; basketball court; benches, a large community garden; a walking path alongside a creek | +ve: Sig. increase in total number of people observed using the park post-intervention (p = 0.004); Increase in proportion of users engaging in moderate (p = 0.007) or vigorous PA (p = 0.04). Post-intervention average monthly visitors sig. increased (p = 0.002) |
Cranney et al. (2016) Sydney, Australia | Quasi-experiment: pre-post time series design | Beachside suburb comprising relatively high socioeconomic status neighborhoods | Outdoor gym installed (60,000 Aus $), targeted marketing and promotional strategies to engage older adults and hosting exercise sessions by a professional Park is 16.08 ha, picnic shelters, barbecues, drinking fountains, toilets and change facilities, a skate park and children’s playground | +ve: Small but sig. increase in senior park users engaging in MVPA at follow-up (1.6 to 5.1%; p < 0.001); sig. increases from baseline to follow-up in the outdoor gym area for MVPA (6–40%; p < 0.001); and seniors’ use (1.4–6%; p < 0.001) |
Slater et al. (2016) Chicago, IL, USA | Quasi-experimental: prospective, controlled, longitudinal design | Predominantly African American and Latino neighborhoods | Park renovations and community engagement (39 intervention parks) Renovations involved replacing old playground equipment and ground surfacing Mean park size 3.86 sq. acres (range 0.09–40.48) | +ve: Sig. increases between baseline and 12-month follow-up for park utilizization and the number of people engaged in MVPA; increase in park utilization over time in intervention parks compared with control |
17.2.3 Greenways and Trail Interventions
Reference | Study design | Population | Intervention | Outcome |
---|---|---|---|---|
Evenson et al. (2005) North Carolina, USA | Quasi-experimental: pre-post design | Adults aged >18 years living within 2 miles of the trail | A railway was converted to a multi-use trail Trail 2.8 miles/10 feet wide with 2 mile spur of 23 mile trail; trail passed by 2 schools, shopping areas, apartment buildings and neighbourhoods |
−ve: Those who had never used the trail had sig. declines in median time spent in MVPA, vigorous PA and bicycling for transport. Those who had used the trail also had sig. declines in median time spent in vigorous PA. |
Burbidge and Goulias (2009) Utah, USA | Quasi-experiment: longitudinal design | Individuals residing near the new trail | Construction of a trail (2-way multi-use trail separated from existing roads and sidewalks) for both transportation and recreation. The trail created a 2.5 mile loop connecting two currently existing sidewalks |
−ve: Negative sig. effect on PA and walking between baseline and follow-up; 18–64 year olds sig. increased number of PA episodes between baseline and follow-up (p = 0.024) |
Fitzhugh et al. (2010) Tennessee, USA | Quasi-experiment: controlled, pre-post design | Children, adolescents and adults in neighbourhood | Retrofit of an urban greenway (2.9 miles long; 8-foot wide) to enhance connectivity of pedestrian infrastructure with nearby retail establishments and schools (cost: $2.1 m) | +ve: Pre and post intervention changes between experimental and control neighbourhoods were sig. different for total PA (p = 0.001); walking (p = 0.001) and cycling (p = 0.038). There was no sig. change over time for active transport to school |
West and Shores (2011) North Carolina, USA | Quasi-experiment: controlled, pre-post design | Residents living within 0.5 mile radius of greenway | 5 miles of greenway developed and added to existing greenway along a river |
−ve: No sig. difference between intervention and control group |
Clark et al. (2014) Southern Nevada, USA | Quasi-experiment: controlled, pre-post design | Trails were in lower SES neighbourhoods | 6 intervention trails: after a marketing campaign promoting PA and trail use (2012), signage was added/altered including: distance markings, way-finding signs, trail maps, trail names, and icons for acceptable uses | +ve: Sig. increases for both control and intervention, pre–post for trail usage per day; 31% increase for the control trails and 35% for the intervention trails (p < 0.01); non-sig. difference between the intervention and control group (p = 0.32) |
Cardiff, Kenilworth and Southampton, United Kingdom | Quasi-experimental, longitudinal design | Adults living within 5 km by road of the core Connect2 projects | Building or improvement of walking and cycling routes across the United Kingdom including a traffic-free bridge over Cardiff Bay; a traffic-free bridge over a busy trunk road; an informal riverside footpath turned into a boardwalk | +ve: Proximity to Connect2 associated with greater use of Connect2; 32% reported using Connect2 at 1 year and 38% at 2 years.; at 2 years, those nearer the intervention sig. increased walking and cycling (15.3 mins/week/km) and total PA (12.5 mins/week/km) |
17.2.4 Greening Interventions
Reference | Study design | Population | Intervention | Outcome |
---|---|---|---|---|
Branas et al. (2011) Philadelphia, PA, USA | Quasi-experiment: difference-in-difference design | Cohort of 50,000 Philadelphians from household survey | Greening of vacant urban land (n = 4436); (> 725,000 m2) from 1999 to 2008 involving removing trash and debris, grading the land, planting grass and trees, installing low wooden fences around perimeter | +ve: Greening associated with reductions in gun assaults (p < 0.001), vandalism (p < 0.001), residents reported less stress and more exercise (p < 0.01) |
Garvin et al. (2013) Philadelphia, PA, USA | Pilot RCT: difference-in-difference analytical approach | People living approx. two blocks surrounding the randomly selected vacant lots; 97% African–American; median income $15,417–17,743 | Greening of vacant lots (4500–5500 square feet); removing debris, grading the land and adding topsoil, planting grass and trees, building a wooden fence | +ve: Non-sig. decrease in the number of total crimes and gun assaults around greened vacant lots compared with control; people around the intervention lots reported feeling sig. safer after greening compared with control lots (p < 0.01) |
Anderson et al. (2014) Cape Town, South Africa | Quasi-experimental, controlled (post data only) | Spectrum of socioeconomic neighbourhoods, ranging from middle to lower income areas | Civic-led greening interventions implemented via three sites | +ive: Biodiversity in the greening intervention sites was higher than the vacant lot and comparable to the conservation sites |
South et al. (2015) Philadelphia, PA, USA | Quasi-experimental, controlled, pre and post | N = 12 participants completed pre- and post-intervention walks; all were African-American, eight male; majority had household income < S15, 000 | Randomly selected cluster of vacant lots received standard greening treatment involving cleaning and removing debris, planting grass and trees, and installing a low wooden post-and-rail fence | +ve: Difference-in-difference estimates between greened and non-greened vacant lots was sig. lower for heart rate (p < .001) for the greened site; being in view of a greened vacant lot decreased heart rate sig. more than a non-greened vacant lot |
Strohbach et al. (2013) Boston, MA, USA | Quasi-experimental, controlled (post data only) | Low SES areas; 617,594 inhabitants; population density of 4939 inhabitants per km2; tree canopy covers 29% of the city area | 12 community driven greening projects in low SES areas including creation of a small park (424 m2),tree plantings in an existing park (4377 m2) and tree plantings at residential houses (859 m2) | +ve: Sig. difference between greening projects and random urban sites (p = .049); most greening projects had more species than the random urban sites in their vicinity |
Jin et al. (2014) Shanghai, China | Quasi-experimental, controlled (post data only) | Area of 6340.5 km2, 23.5 million population | Street trees on 6 streets (length 205–223 m; width 15.2–17.5 m) were treated with different pruning intensities (strong, weak and null) which would result in different canopy coverage across the four seasons | +ve: Increased street tree canopy was positively associated with PM2.5 concentrations owing to reduced air circulation |
Ward Thompson et al. (2014) England, Scotland and Wales, United Kingdom | Quasi-experiment: controlled, pre-post design | Mean age 75 years; 44% male; 22.5% non-white British | n = 56 residents pre and n = 29 post intervention ‘DIY Streets’: 9 intervention streets located in urban areas in United Kingdom. Streets were made safer, more attractive and traffic calming measures were added. | +ve: Sig. positive perceptions of intervention streets post-intervention (p = 0.04); longitudinal participants perceived they were sig. more active post-intervention (p = 0.04) than the control group |
Joo and Kwon (2015) Suwon, South Korea | Quasi-experimental, controlled (post data only) | Population 1.2 m | 74 sites with street greenery (e.g. planter boxes) installed by the city council, located in low-rise residential areas to reduce illegal dumping of household garbage | +ve: Illegal dumping of household garbage occurred at 55.4% of sites with installed greenery compared to 91.9% of sites without greenery installed |
17.2.5 Green Infrastructure Interventions
Reference | Study design | Intervention | Outcome |
---|---|---|---|
Van Seters et al. (2009) Toronto, Canada | Quasi-experiment, controlled (post data only) | A 241 m2 green roof vegetated with wildflowers installed on a multi-story, university building | +ve: The green roof retained 63% more rainfall than the conventional roof over the 18 month monitoring period |
Carpenter and Kaluvakolanu (2011) Michigan, USA | Quasi-experiment, controlled (post data only) | Extensive green roof of 10.16 cm depth applied to the roof of a building on a university campus; a green roof section of 325.2 m2 and 929 m2 were monitored | +ve: Sig. higher total solids concentration (p = 0.045) for the green roof than the asphalt roof; lower total phosphate concentrations for the green roof (non-sig.); green roof retained 68% of rainfall volume and reduced peak discharge by an average of 89% |
Mayer et al. (2012) Ohio, USA | Before-after-control-intervention (BACI) experimental design | Retro-fit storm water management: Installation of 83 rain gardens and 176 rain barrels onto more than 30% of the 350 eligible residential properties through an incentivised auction (2007–2008) | +ve: Intervention had an overall small but sig. effect of decreasing storm water quantity at the sub watershed scale |
Fassman-Beck et al. (2013) Auckland, New Zealand | Quasi-experiment, controlled (post data only) | A 500 m2 extensive green roof installed on a council civic centre | +ve: 57% retention of rain water in comparison to control |
Ohio, USA | Before–after–control–intervention (BACI) experimental design | Retro-fit storm water management: Installation of 81 rain gardens and 165 rain barrels onto 30% of properties through an incentivised auction (2007–2008) at 4 experimental subcatchments |
−ve: No sig. difference between control and experimental sites with regards to stream water quality, periphyton, and macroinvertebrate metrics +ve: Small sig. decrease in runoff volume in treatment subcatchments |
Kondo et al. (2015) Philadelphia, PA, USA | Quasi-experiment: difference-in-difference design | Installation of green storm water infrastructure at 52 sites: 152 tree trenches, 46 infiltration or storage trenches, 43 rain gardens, 29 pervious pavement installments, 20 bumpouts, 14 bio-swales, 5 storm water basins, 1 wetland, and 12 other | +ve: Sig. reductions in narcotics possession (18–27% less) (p < .01), (p < .01) at varying distances from treatment sites; sig. reductions in narcotics manufacture and burglaries; non-sig. reductions in homicides, assaults, thefts, public drunkenness, stress levels, blood pressure and cholesterol |
Jarden and Jefferson (2016) Ohio, USA | Before–after–control–intervention (BACI) experimental design | Installation of 91 rain gardens, street-connected bio-retention cells and rain barrels at 2 treatment streets. Rain gardens (< 25 m2) were installed in front yards and backyards; bio-retention cells (~26–44 m2) were installed between the sidewalk and street | +ve: Reduction in storm water flow at the treatment streets with reductions of up to 33% of peak discharge and 40% of total run-off volume |
Peng and Jim (2015) Hong Kong, China | Quasi-experiment, controlled, pre and post design | A 484 m2 extensive green roof was retrofitted on a 2-story railway station | +ve: Green roof displayed cooling effects in spring, summer, and fall, with slight warming effects in winter |