Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Aug 1.
Published in final edited form as: Urban For Urban Green. 2023 Jun 28;86:128015. doi: 10.1016/j.ufug.2023.128015

Association between urban greenspace, tree canopy cover and intentional deaths: An exploratory geospatial analysis

Sophia C Ryan 1, Margaret M Sugg 1, Jennifer D Runkle 2
PMCID: PMC10486201  NIHMSID: NIHMS1914302  PMID: 37693117

Abstract

Greenspaces can provide restorative experiences, offer opportunities for outdoor recreation, and reduce mental fatigue; all of which may improve community health and safety. Yet few studies have examined the neighborhood-level benefits of greenspace in reducing violent deaths. This study explored the association between three distinct greenspace metrics: public greenspace quantity, public greenspace accessibility, neighborhood tree canopy cover, and intentional deaths (i.e., homicides and suicides). Generalized linear models and spatial error models investigated the association between greenspace, tree canopy and intentional deaths in three geographically distinct cities in North Carolina. Results revealed that increased neighborhood greenspace accessibility and tree canopy cover were associated with reduced intentional deaths in all three urban areas. Neighborhood greenspace accessibility was the most protective factor across all study areas. The relationship between neighborhood greenspace accessibility and intentional deaths was more significant for non-firearm deaths as compared to firearm deaths, indicating that weapon type may be an important consideration for neighborhood greenspace interventions. Compared to predominantly White neighborhoods, predominantly Black neighborhoods had higher rates of homicide in Asheville and Durham and higher rates of suicide in Charlotte. Future policy and research should focus on improving equitable access to existing and future greenspaces, especially in primarily Black neighborhoods.

Keywords: Suicide, Homicide, Spatial Regression, greenspace accessibility, greenspace quantity, green space

1. Introduction

Both greenspaces and tree canopy cover have been linked to numerous community benefits, including protective effects for human health (El-Mallakh et al., 2022; Lee et al., 2023) reductions in crime (Gilstad-Hayden et al., 2015; Troy et al., 2012), and intentional deaths (i.e., homicide and suicide) (El-Mallakh et al., 2022). Greenspaces are thought to provide restorative experiences, contribute to stress reduction, offer avenues for social cohesion, and provide opportunities for physical recreation (Markevych et al., 2017; Wang et al., 2021), all of which may improve population mental health.

Worldwide, males are the most common perpetrator and victims of homicide, though women bear the greatest burden of intimate partner violence and death at the hands of a partner or family member (UNODC, 2019). Males are also more likely to die by suicide (AFSP, 2022) and are less likely to seek mental health care or take preventative measures (Parent et al., 2018). In the United States, firearms are the most common weapon associated with homicides and suicides (UNODC, 2019), stressing the need for research investigating low-cost community interventions that can help reduce violent deaths, with additional consideration of the mechanism of death (weapon used). Evidence suggests that both the built and social environments are strongly associated with both homicide and suicide rates; with a lack of social cohesion and income inequality predicting high rates of both homicide and suicide at the community-level (Wilkins et al., 2019).

Recent research suggests homicide rates are lower in and around parks and greenspaces (Breetzke et al., 2020). Urban greenspace has been associated with a reduction in criminal behavior and a corresponding decline in violent crimes as greenspaces in these areas may promote more social gatherings and community cohesion, reduce the urban heat island effect, and promote positive community perception around place (Shepley et al., 2019). Furthermore, greenspaces and natural features are thought to reduce mental fatigue and distress, potentially a precursor to many violent crimes (Gilstad-Hayden et al., 2015; Kuo and Sullivan, 2001).

Increased greenspace quantity is associated with population-level decreases in poor mental health outcomes (decreases in community mental health burdens), including suicide (Helbich et al., 2018; Jiang et al., 2021). Among greenspace-mental health research, the definition of accessibility and quantity of greenspace has varied, with some studies identifying public, usable greenspaces (Browning and Rigolon, 2018; Houlden et al., 2019; Mears and Brindley, 2019), while others consider land cover and NDVI as indicators of the quantity of greenspace (Beyer et al., 2014; Sarkar et al., 2018; Wang et al., 2021). Despite differences in how the amount of greenspace is operationalized across studies, the evidence shows that poor mental health outcomes decrease as greenspace quantity increases. Findings further suggest that decreased distance to or increased accessibility of green space may all be associated with mental health benefits (Shen and Lung, 2018).

The accessibility of greenspaces may also influence the greenspace-community health relationship. Buffers, specified areas surrounding features of interest, are commonly utilized to ascertain accessibility and can range from 100 meters to 3 kilometers or larger (Breetzke et al., 2020; Mears and Brindley, 2019; Nutsford et al., 2013). When determining accessibility, most studies consider walkability, which typically translates to an 800-meter buffer (Ekkel and de Vries, 2017; Houlden et al., 2019). Findings suggest that regardless of the accessibility buffer distance, more access translates to better mental health outcomes (Ekkel and de Vries, 2017; Houlden et al., 2019). While buffer analysis provides important information regarding the amount of greenspace accessible within a certain radius; more precise estimates of accessibility (e.g., distance to nearest greenspace) may provide new knowledge regarding the association between greenspace accessibility and intentional deaths. Results can help inform future greenspace developments as to whether access to greenspace or the total quantity of greenspace is most important for overall community health.

High level of tree canopy cover (a metric that accounts for the amount of trees in a given location) has also been linked with lower rates of violent crime, property crime, and all crime (Gilstad-Hayden et al., 2015; Troy et al., 2012). After controlling for other community-level variables, including educational attainment, socioeconomic status, and housing vacancies, past research investigating tree cover and crime has found significant, substantial decreases in crimes as tree cover increases, with a 10% increase in tree cover associated with up to a 12–15% decrease in crime (Gilstad-Hayden et al., 2015; Troy et al., 2012). In addition, Lee et al., (2023) showed that as tree canopy cover increases, suicide attempts decrease; suggesting trees have wide-reaching community-health promotion and safety benefits.

Yet, the evidence of greenspace as a protective intervention for violent deaths is inconclusive (Bogar and Beyer, 2016; Shepley et al., 2019; Sreetheran and Bosch, 2014). While many studies have found greenspace to benefit mental health (Collins et al., 2020); the perceived danger of public parks can inhibit visitation and create the illusion, whether true or untrue, that parks are dangerous and hot spots for crime (Sreetheran and Bosch, 2014). The relationship between greenspace and intentional deaths (i.e., homicide and suicide) elicits further consideration, as social cohesion opportunities provided by greenspaces may foster more connected communities and improve population-level mental health (Bogar and Beyer, 2016). However, if greenspaces are perceived as dangerous due to high crime rates, many community members may not find the benefits of greenspace worth the risk of access (Bogar and Beyer, 2016). As such, further investigation into the relationships between multiple greenspace metrics (i.e., tree cover, public greenspace quantity, public greenspace accessibility) and intentional deaths is needed to help guide future community health interventions. Consideration of multiple metrics will provide important information on whether active interaction with public greenspace (greenspace quality and accessibility) or passive interaction with greenspace (tree cover) are protective of suicide or homicide related deaths in urban areas.

Despite mounting evidence linking greenspace to community safety and individual-level mental health benefits, little research has investigated the relationship between greenspaces and intentional deaths (i.e., suicides and homicides) (El-Mallakh et al., 2022; Sadatsafavi et al., 2022). Moreover, at the time of publication, no research has been conducted in North Carolina, a state in the Southeastern US. To further investigate this relationship, our analysis considered the relationship between intentional deaths, greenspace access, and tree cover in three major cities in North Carolina, USA, at the neighborhood level. Secondary analysis included weapon-type as an indicator of the manner of intentional death to assess if greenspace was associated with specific manners of death. Results can guide future research and public policy regarding the development of natural spaces for community health and safety interventions.

2. Methods

2.1. Study Area

This analysis was conducted at the neighborhood-level in three cities in North Carolina; Asheville (population 94,067), Charlotte (population 879,709), and Durham (population 285,527) (U.S. Census Bureau, 2022) (Figure 1). For this analysis, neighborhoods were defined as census block groups (Duncan and Kawachi, 2018). Asheville, is located in the western region of the state and is surrounded by National Forest and Game lands, in addition to having greenspace access in the form of public parks throughout the city. Asheville is 81% white, 10% Black, 7% Hispanic or Latino; 12% of individuals live in poverty (U.S. Census Bureau, 2022), Asheville is an urban center in a predominantly rural region with substantial greenspace access. Durham is located in the urban center, ‘the triangle,’ comprising Raleigh, Durham, and Chapel Hill. Durham has greater racial diversity than Asheville; 47% of residents are white, 37% are Black, and 13% are Hispanic or Latino; 13.5% of residents are living in poverty (U.S. Census Bureau, 2022). Durham has ample greenspace predominantly in the form of public parks and was selected due to its high count of homicides. Charlotte is located in the south of the state and is one of the most rapidly expanding urban centers in the United States (Kenan Institute, 2022). Charlotte, as with Durham, has high racial diversity; 44% of residents are white, 35% are Black and 15% are Hispanic or Latino; 11% of individuals are living in poverty. Greenspace in Charlotte is primarily in the form of public parks. Charlotte was selected for this study because it has the state’s highest number of intentional deaths. We elected to conduct this analysis in three separate cities characterized by distinct differences in available public greenspaces, demographics, and intentional death counts to better understand the relationship between intentional deaths, greenspace, and tree cover in urban areas.

Figure 1 -.

Figure 1 -

Open in a new tab

Map of (A) North Carolina, indicating the three cities considered in this analysis, (B) Durham county; which includes the city of Durham, neighborhood boundaries (census block groups), and public greenspace, (C) Buncombe county; which includes the city of Asheville, neighborhood boundaries (census block groups) and public greenspace, and (D) Mecklenburg county; which includes the city of Charlotte, neighborhood boundaries (census block groups) and public greenspace.

2.2. Data

2.2.1. Mortality Data

Intentional death data for suicides and homicides from 2004–19 were obtained for the state of North Carolina (NC) from the National Violent Death Reporting System (NC-VDRS). NC-VDRS provides comprehensive reporting of all violent deaths and contains additional contextual factors (e.g., race, age, manner of death, history of mental health) (NCDHHS, 2022). Individual manner of death is determined by abstractor review based on information from the death certificate, law enforcement, and medical examiner records (NCDHHS, 2022). Deaths are categorized as suicide, homicide, unintentional firearm, undetermined intent, and legal intervention (NCDHHS, 2022). For this analysis, data were restricted to only suicides and homicides occurring within the county boundaries of Asheville (Buncombe County), Durham (Durham County), and Charlotte (Mecklenburg County).

NC-VDRS data provides the deceased’s zip code, county, and state of residence but does not provide the individual’s address. In order to determine the deceased’s residential address, death certificate data were obtained for the same time period (2004–19) from the North Carolina Department of Health and Human Services (NCDHHS) (NCDHHS, 2022). Probabilistic joins matched the NC-VDRS death data with the corresponding NCDHHS death certificate data. All death certificate joins had a success rate of 90% or higher. Joins were completed in RStudio 2022.02.3 (RStudio Team, 2022). Addresses were then geocoded in ArcGIS Pro 2.9.0 (ESRI, 2022) to provide the residential location of all individuals.

Our analysis investigated the greenspace-intentional death relationship at the community-level; therefore, we adjusted for population by operationalizing deaths as deaths per 100,000 people per census block group (Table 1). The census block group was elected as the geographic unit of analysis as it is the finest spatial scale with available community-level information.

Table 1.

Breakdown of variables considered in statistical models. Greenspace data is from PAD_US and ParkServe; death data is from the North Carolina Violent Death Reporting System (2004–19); covariate data is from American Community Survey (2018 5-year estimates).

Variables
Deaths Intentional Deaths/100,000 people per census block group
Suicides/100,000 people per census block group
Homicides/100,000 people per census block group
Greenspace Percent greenspace: Public lands/census block group
Tree Canopy: Average % tree canopy cover (amount of trees across all land types)/census block group
Greenspace Accessibility: Average distance from residential location to nearest greenspace per census block group
Covariates ICE Metrics: Economic and racial extremes, measured at the census block group
Open in a new tab

2.2.2. Greenspace Data

Public greenspaces have been associated with protective effects for mental health (Collins et al., 2020; Houlden et al., 2021) and reduced suicide rates (Jiang et al., 2021); however, less attention has been directed at the relationship between public greenspace and homicides. Tree canopy cover was considered as past research has demonstrated that increased tree canopy cover is associated with reduced violent crimes (Gilstad-Hayden et al., 2015; Troy et al., 2012), and both homicides and suicides at the community-level (El-Mallakh et al., 2022). For our analysis, greenspace was operationalized as three separate metrics (1) Public Greenspace Accessibility, or the average distance from each residential location to a public greenspace location averaged across each census block group; (2) Percent Public Greenspace, public greenspace per census block group; and (3) Neighborhood Tree Canopy, or total greenspace (both public and private lands) per census block group (SM Table 1).

Greenspaces were identified using the Protected Area Database of the United States (PAD-US) (U.S. Geological Survey (USGS) Gap Analysis Project (GAP), 2020) and the Trust for Public Land’s ParkServe dataset (The Trust for Public Land, 2021) (Figure 1). PAD-US is a spatial dataset of all government-managed greenspaces (e.g., national forest land, national parks, and historical areas). To ensure selected greenspaces were publicly accessible, selection criteria were applied to remove restricted access to public lands (e.g., military bases) (Browning et al., 2022). ParkServe is a dataset comprising all public parks (e.g., local and city parks) (The Trust for Public Land, 2021). No additional selection criteria were applied to the ParkServe dataset. Percent Public Greenspace was calculated from these two datasets as the percent public greenspace per census block group. Whereas, Public Greenspace Accessibility, was determined by calculating the distance between the residential address and the nearest greenspace using the ‘near’ tool in ArcGIS Pro 2.9.0 to determine accessibility metrics (ESRI, 2022). Neighborhood Tree Canopy was calculated using the United States Forest Service 30-meter resolution tree canopy cover dataset (USFS, 2019). Tree Canopy data from 2016 captures tree canopy cover on both public and private lands. Percent calculations were performed using the tool ‘zonal statistics as table’ in ArcGIS Pro 2.9.0 (ESRI, 2022). Tree canopy cover is reported as a percentage, ranging from 0%, with no trees present, to 100%, indicating full tree canopy coverage (Table 1).

2.2.3. Covariates

Analyses were adjusted for census block group race and socio-economic status using the Index of the Concentration of Extremes (ICE) (Krieger et al., 2016), as intentional deaths can vary based on race and socioeconomic status (Beard et al., 2017; Wilkins et al., 2019). ICE is an index that considers extreme concentrations of privilege and deprivation by analyzing the spatial distribution of income and race using US Census Data (Krieger et al., 2016). For this analysis, two metrics were considered. The first metric, ICE: Income, measures community income extremes by comparing how many households make over $100,000 per year and under $25,000 per year using American Community Survey 2018 5-year estimates (US Census, 2018). The second metric, ICE: Race, communicates the racial demographics of a community by comparing the number of non-Hispanic Black residents to the number of non-Hispanic white residents. The ICE metrics range from −1 (least privilege) to +1 (most privilege) (Krieger et al., 2016). For this analysis, ICE metrics were computed as tertiles (T1: most deprived and T3: least deprived). The use of tertiles was adapted to improve the interpretability of the regression results; where Tertile 1 corresponds to predominately low income (ICE: Income) and predominately Black (ICE: Race); Tertile 2 corresponds to mixed income (ICE: Income) and mixed race (ICE: Race), and Tertile 3 corresponds to predominantly high income (ICE: Income) and predominantly White (ICE: Race).

2.3. Statistical Analysis

Census-block analyses leveraged two statistical methods to investigate the association between intentional deaths and greenspace; generalized linear models (GLMs) and spatial error models. GLMs with log-corrected variables and gaussian distributions were run to initially investigate the relationship between intentional deaths and greenspace metrics. The initial analysis considered aggregated deaths from all three cities in one model; whereby three separate models were run for intentional deaths, suicides, and homicides. Next, three GLMs were run for each city separately, one each for all intentional deaths, suicides, and homicides. GLMs adjusted for community-level race and socio-economic status using the ICE: Race and ICE: Income metrics. Every model considered all three greenspace metrics. Each model tested for multicollinearity by calculating the variance inflation factor (VIF) in RStudio 2022.02.3 (RStudio Team, 2022). No variables exceeded the threshold of VIF=2; therefore none were removed from the analysis for violating the assumption of independence (Craney and Surles, 2002; James et al., 2013).

All models were tested for spatial autocorrelation using Moran’s I. Moran’s I p-values at or below 0.05 indicate spatial autocorrelation and indicate spatial dependence pointing to the need to perform spatial regression (Anselin et al., 2008; Legendre, 1993). Spatial autocorrelation was present in Charlotte and Durham but was not identified for Asheville or the aggregated three-city model. In the absence of significant spatial autocorrelation, spatial regression models are not required (i.e., Asheville and three-city model) (Anselin et al., 2008). Using the Lagrange multiplier diagnostics for the spatial dependence test, the spatial error model was selected as the best spatial regression model for Charlotte (LM =0.35) and Durham (LM = 4.62). As with the GLMs, spatial error models were run using log-transformed death counts and a Gaussian distribution. As sub-an additional supplemental analysis, we elected to stratify deaths by weapon type in the aggregated city model. We categorized weapons as firearms or non-firearms as most deaths (62.253%–65.4%) resulted from the use of firearms (Table 2).

Table 2.

Demographic characteristics of individuals who died by suicide and homicide in Asheville, Charlotte and Durham, NC, 2004–19. Data was obtained from the NC VDRS. Percentages represent the percent of total violent deaths by place and demographic variables.

Asheville n (%) Charlotte n (%) Durham n (%) Total n (%)
Suicide (n) 602 (26.9) 1,299 (58.0) 338 (15.9) 2,239
Homicide (n) 125 (8.3) 962 (63.7) 423 (28.0) 1,510
Total Deaths (n) 727 (19.4) 2,261 (60.3) 761 (20.3) 3,749
Average Age, years 46.97 39.2 38.29 -
Race (%)
Indigenous 2 (0.3) 6 (0.3) 3 (0.4) 9 (0.2)
Asian 9 (1.2) 45 (2.0) 12 (1.6) 66 (1.8)
Black 58 (8.0) 963 (42.6) 414 (54.4) 1,435 (38.3)
White 657 (90.4) 1,205 (53.3) 324 (42.6) 2,186 (58.3)
Other 1 (0.1) 42 (1.9) 8 (1.0) 51 (1.4)
Sex (%)
Male 551 (75.8) 1786 (79.0) 606 (79.6) 2,943 (78.5)
Female 176 (24.2) 475 (21.0) 155 (20.4) 806 (21.5)
Weapon Type(%)
Firearm 385 (53.0) 1448 (64.0) 498 (65.4) 2,331 (62.2)
Sharp Instrument 37 (5.1) 126 (5.6) 62 (8.1) 225 (6.0)
Poisoning 117 (16.1) 224 (9.9) 68 (8.9) 409 (10.9)
Hanging+ 134 (18.4) 334 (14.8) 88 (11.6) 556 (14.8)
Other* 54 (7.4) 129 (5.7) 45 (6.0) 228 (6.1)
Open in a new tab
+

Hanging includes deaths from strangulation and suffocation

*

Other includes: Non-powder gun, blunt instrument, personal weapon, fall, fire/burns, shaking, vehicles, intentional neglect, and unknown.

As a sensitivity analysis, models were stratified by season, where deaths were categorized as occurring in the warm season (April-October) and cold season (November-March) to determine if seasonality influenced the association between violent deaths and greenspace metrics. Results showed that the relationships did not vary significantly between warm and cold-season deaths. These results are summarized in SM Table 1.

3. Results

Table 2 summarizes the demographic characteristics of all intentional deaths in Asheville, Charlotte and Durham from 2004–2019. Asheville had the fewest intentional deaths (n=727), Charlotte had the greatest number of intentional deaths (n=2,261). Durham had the highest proportion of homicides (n=423). In Charlotte and Asheville, deaths were highest among White individuals (53.3–90.4%), and in Durham deaths were highest among Black individuals (45.4%). In all three cities, deaths were highest for males (75.8–79.6%). Asheville had the highest average age at the time of death (47 years) compared to the other study locations. Of all three study areas, Asheville census block groups have the least amount of public greenspace, where 58% have public greenspace (SM Table 2); Durham has the highest percent of census block groups with public greenspace (76%). All three cities had varied tree cover, ranging from 0.16–84.44% in Asheville, 0.15–78.61% in Charlotte, and 2.4–72.63% in Durham. Asheville had the least accessible public greenspace, with distances to nearest greenspace ranging from 0 kilometers to 7.45 kilometers, and Charlotte had the best neighborhood greenspace accessibility, with distances ranging from 0 kilometers to 2.78 kilometers.

3.1. GLMs and Spatial Error

Table 3 summarizes GLM results for all cities combined (Model 1), and Asheville (Model 2), and spatial error model results for Charlotte (Model 3) and Durham (Model 4). Spatial error models were used to adjust for spatial autocorrelation in Charlotte and Durham. The aggregated GLM found that decreased public greenspace accessibility (increasing average distance to nearest greenspace) was significantly associated with higher intentional deaths (0.0005, p<0.001) and suicides (0.0006, p<0.001); the association between accessible greenspace and homicides was not significant (0.00015, p=0.11). Neighborhood tree canopy cover was associated with reduced intentional deaths (−0.004, p=0.18), suicides (−0.002, p=0.57), and homicides (−0.003, p=0.51). In contrast to other greenspace measures, neighborhood greenspace quantity, operationalized as percent greenspace land cover, was significantly associated with higher counts of all intentional deaths (0.027, p<0.001) and suicides (0.031, p<0.001).

Table 3-.

Generalized linear model (Model 1 and Model 2) and spatial error model (Model 3 and Model 4) regression results. Dependent death variables were log-transformed to achieve a normal distribution.

Model 1 - All Cities
Death Suicide Homicide
Estimate (SE) VIF P Estimate (SE) VIF P Estimate (SE) VIF P
Average Distance to Nearest Greenspace (m) 0.0005 (0.00007) 1.27 <0.001 0.0006 (0.0001) 1.27 <0.001 0.00015 (0.00009) 1.27 0.11
Average Tree Canopy Cover −0.004 (0.003) 1.21 0.18 −0.002 (0.004) 1.21 0.57 −0.003 (0.0042) 1.21 0.51
Percent Greenspace 0.027 (0.006) 1.14 <0.001 0.031 (0.007) 1.14 <0.001 0.0093 (0.008) 1.14 0.27
ICE Race(Primarily Black) 0.68 (0.13) 1.42 <0.001 −0.55 (0.16) 1.42 <0.001 2.59 (0.18) 1.42 <0.001
ICE Race(Mixed Race) 0.10 (0.12) 1.42 0.4 −0.10 (0.14) 1.42 0.5 0.868 (0.166) 1.42 <0.001
ICE Race (Primarily White) REFERENCE
ICE Income(Primarily Low) 0.52 (0.12) 1.31 <0.001 0.271 (0.155) 1.31 0.08 0.88 (0.18) 1.31 <0.001
ICE Income(Mixed) 0.25 (0.17) 1.31 0.03 0.366 (0.146) 1.31 0.01 0.223 (0.17) 1.31 <0.001
ICE Income (Primarily High) REFERENCE
R2 0.146 0.105 0.315
Model 2 - Asheville
Death Suicide Homicide
Estimate (SE) VIF P Estimate (SE) VIF P Estimate (SE) VIF P
Average Distance to Nearest Greenspace (m) 0.0002 (0.001) 1.5 0.06 0.0002 (0.0001) 1.5 0.05 0.0002 (0.0002) 1.5 0.32
Average Tree Canopy Cover −0.003 (0.005) 1.32 0.59 −0.004 (0.005) 1.32 0.45 −0.017 (0.01) 1.32 0.1
Percent Greenspace 0.0133 (0.01) 1.28 0.19 0.016 (0.01) 1.28 0.13 0.032 (0.02) 1.27 0.11
ICE Race (Primarily Black) 0.48 (0.26) 1.45 0.06 0.41 (0.27) 1.45 0.13 0.98 (0.51) 1.45 0.06
ICE Race (Mixed Race) 0.022 (0.24) 1.45 0.93 0.12 (0.25) 1.45 0.64 −0.33 (0.49) 1.45 0.5
ICE Race (Primarily White) REFERENCE
ICE Income (Primarily Low) −0.146 (0.25) 1.23 0.55 −0.156 (0.258) 1.23 0.55 −0.05 (0.50) 1.23 0.93
ICE Income (Mixed) −0.019 (0.23) 1.23 0.93 −0.168 (0.238) 1.23 0.48 0.27 (0.45) 1.23 0.55
ICE Race (Primarily High) REFERENCE
R2 0.06 0.043 0.085
Model 3 - Charlotte
Death Suicide Homicide
Estimate (SE) VIF P Estimate (SE) VIF P Estimate (SE) VIF P
Average Distance to Nearest Greenspace (m) 0.0015 (0.0001) 1.2 <0.001 0.002 (0.0002) 1.21 <0.001 0.001 (0.0002) 1.18 0.02
Average Tree Canopy Cover −0.01 (0.005) 1.15 0.04 −0.0028 (0.006) 1.15 0.61 −0.0013 (0.006) 1.14 0.83
Percent Greenspace 0.053 (0.008) 1.18 <0.001 0.06 (0.013) 1.19 <0.001 0.0138 (0.013) 1.18 0.27
ICE Race (Primarily Black) −0.276 (0.152) 1.95 0.07 0.44 (0.201) 1.88 0.03 −0.989 (0.23) 1.72 <0.001
ICE Race (Mixed Race) 0.11 (0.19) 1.95 0.57 0.732 (0.25) 1.88 0.004 −1.816 (0.28) 1.72 <0.001
ICE Race (Primarily White) REFERENCE
ICE Income (Primarily Low) −0.65 (0.15) 1.94 <0.001 0.307 (0.203) 1.87 0.13 −1.27 (0.23) 1.72 <0.001
ICE Income (Mixed) −1.48 (0.19) 1.94 <0.001 −0.46 (0.25) 1.87 0.07 −2.34 (0.28) 1.72 <0.001
ICE Race (Primarily High) REFERENCE
R2 0.27 0.16 0.37
Model 4 - Durham
Death Suicide Homicide
Estimate (SE) VIF P Estimate (SE) VIF P Estimate (SE) VIF P
Average Distance to Nearest Greenspace (m) 0.001 (0.0002) 1.34 <0.001 0.0011 (0.0003) 1.32 <0.001 0.00004 (0.0003) 1.37 0.88
Average Tree Canopy Cover −0.005 (0.008) 1.41 0.52 0.002 (0.011) 1.38 0.86 −0.011 (0.01) 1.44 0.27
Percent Greenspace 0.012 (0.014) 1.24 0.39 0.017 (0.02) 1.24 0.38 −0.018 (0.019) 1.21 0.34
ICE Race (Primarily Black) 1.61 (0.31) 1.43 <0.001 −0.028 (0.43) 1.41 0.95 2.68 (0.40) 1.41 <0.001
ICE Race (Mixed Race) 0.286 (0.283) 1.43 0.31 0.202 (0.4) 1.41 0.61 0.34 (0.0.37) 1.41 0.37
ICE Race (Primarily White) REFERENCE
ICE Income (Primarily Low) −0.456 (0.29) 1.22 0.12 −0.41 (0.4) 1.21 0.31 −0.22 (0.39) 1.21 0.57
ICE Income (Mixed) −0.588 (0.28) 1.22 0.03 −0.48 (0.38) 1.21 0.21 −0.71 (0.36) 1.21 0.05
ICE Race (Primarily High) REFERENCE
R2 0.27 0.11 0.35
Open in a new tab

Predominately Black neighborhoods were associated with higher rates of intentional death (0.68, p<0.001) and homicide (2.59, p<0.001), and lower rates of suicide (−0.55, p<0.001). Low and mixed-income neighborhoods were associated with slightly higher rates of homicide (0.88, p<0.001) and suicide (0.366, p=0.01). The weapon type analysis revealed that neighborhood tree canopy cover was significantly associated with lower rates of suicides and homicides, when the deceased was killed with a non-firearm weapon. However, when looking at only firearms, this relationship did not remain significant (Table 4).

Table 4 -.

Stratified weapon regression results. Deaths were categorized as (1) firearm or (2) non-firearm. Data is from the North Carolina Violent Death Reporting System, 2004–19.

Firearms Homicide Suicide Death
Estimate (SE) P Estimate (SE) P Estimate (SE) P
Average Tree Canopy Cover −0.0048 (0.004) 0.24 0.0025 (0.004) 0.54 −0.002 (0.004) 0.58
Percent Greenspace 0.019 (0.01) 0.02 0.048 (0.01) <0.001 0.041 (0.01) <0.001
Average Distance to Nearest Greenspace (m) 0.0003 (0.0001) 0.001 0.001 (0.0001) <0.001 0.001 (0.0001) <0.001
ICE Race (Primarily Black) 2.77 (0.17) <0.001 −0.34 (0.17) 0.05 1.34 (0.15) <0.001
ICE Race (Mixed Race) 0.99 (0.16) <0.001 0.071 (0.16) 0.66 0.44 (0.14) 0.002
ICE Income (Primarily Low) 0.75 (0.17) <0.001 0.18 (0.17) 0.29 0.56 (0.15) <0.001
ICE Income (Mixed) 0.007 (0.16) 0.67 0.28 (0.16) 0.08 0.19 (0.14) 0.18
Non-Firearms Homicide Suicide Death
Estimate (SE) P Estimate (SE) P Estimate (SE) P
Average Tree Canopy Cover −0.0043 (0.0035) 0.22 −0.013 (0.004) 0.002 −0.015 (0.004) <0.001
Percent Greenspace −0.0042 (0.0069) 0.55 0.043 (0.008) <0.001 0.039 (0.008) <0.001
Average Distance to Nearest Greenspace (m) 0.00028 (0.00001) <0.001 0.0011 (0.00009) <0.001 0.0012 (0.00009) <0.001
ICE Race (Primarily Black) 0.89 (0.15) <0.001 −0.35 (0.17) 0.04 0.231 (0.17) 0.16
ICE Race (Mixed Race) 0.29 (0.14) 0.04 0.12 (0.16) 0.46 0.228 (0.15) 0.14
ICE Income (Primarily Low) 0.554 (0.15) 0.002 −0.11 (0.17) 0.52 0.134 (0.16) 0.41
ICE Income (Mixed) −0.02 (0.14) 0.89 0.17 (0.16) 0.29 0.08 (0.15) 0.6
Open in a new tab
**

High ICE Metrics are reference (Primarily White & Primarily High Income)

As with the aggregated, three-city model, GLMs in Asheville found that decreased public greenspace accessibility (increased average distance to nearest greenspace) was associated with higher intentional deaths (0.0002, p=0.06) and suicides (0.0002, p=0.05). Neighborhood tree canopy cover was also associated with lower deaths; however, this relationship was not significant. Increased percent greenspace land cover was associated with more intentional deaths, though this relationship was not significant. In Asheville, predominantly Black neighborhoods were associated with higher homicide death counts (0.98, p=0.06), as compared to predominantly White neighborhoods.

In Charlotte, spatial error models found decreasing public greenspace accessibility was significantly associated with higher intentional deaths (0.0015, p<0.001), suicides (0.002, p<0.001) and homicides (0.001, p<0.001), such that, as average distance to nearest greenspace increased, intentional deaths also increased. Tree canopy cover was significantly associated with lower intentional deaths (−0.01, p=0.04), though this relationship was not significant when considering only suicide and only homicide. Percent greenspace was significantly associated with a higher number of intentional deaths (0.053, p<0.001) and suicides (0.06, p<0.001). Predominately Black and mixed-race neighborhoods were associated with higher rates of suicides (0.732, p=0.004) and lower rates of homicides (−1.816, p<0.001), compared to predominately White neighborhoods. Low and mixed-income neighborhoods were associated with fewer intentional deaths (−1.48, p<0.001) and homicides (−2.34, p<0.001), as compared to high income neighborhoods.

In Durham, spatial error models also found that all intentional deaths (0.001, p<0.001) and suicides (0.0011, p<0.001) increased as distance to nearest greenspace increased (i.e. accessibility decreased). Increased neighborhood tree canopy cover was associated with lower intentional deaths and homicides, though this relationship was not significant. As with Charlotte and Asheville, increased greenspace quantity was associated with a higher number of intentional deaths and suicides; however, increased greenspace quantity was associated with a lower number of homicides, though this relationship was not significant. Predominantly Black neighborhoods were associated with higher rates of intentional death (1.61, p<0.001) and homicide (2.68, p<0.001), compared with predominantly White neighborhoods. Mixed-income neighborhoods were associated with fewer homicides (−0.71, p=0.05), as compared to high income neighborhoods.

4. Discussion

This analysis investigated the associations between neighborhood greenspace quantity, accessibility, and tree canopy cover and intentional deaths in three distinct cities in North Carolina. Results revealed that across all three cities the number of intentional deaths increased as average distance to nearest greenspace increased (i.e., accessibility decreased), suggesting that high greenspace accessibility may be protective against intentional deaths. The abundance of tree canopy cover was associated with lower intentional deaths, particularly in Charlotte (i.e., a city with a high rate of intentional deaths), corroborating findings from past studies which have found that trees may be protective against suicides and homicides (El-Mallakh et al., 2022; Gilstad-Hayden et al., 2015; Lee et al., 2023; Troy et al., 2012). Interestingly, we observed that both high neighborhood tree canopy cover and greenspace accessibility were associated with a lower risk of non-firearm related deaths; firearm deaths were only significantly associated with greenspace accessibility. Our study is one of the first to consider weapon types. Results indicate that means reduction (reducing a suicidal person’s access to highly lethal means) may be an important consideration for neighborhood greenspace interventions.

Few studies have considered the protective effect of greenspace distance on lower risk of intentional deaths. Our results suggest that living in urban areas with less public greenspace accessibility is associated with higher rates of intentional deaths, namely suicide in all three cities and homicide in Charlotte, after controlling for neighborhood race and socioeconomic status. Past analyses have used buffers to investigate the role greenspace accessibility plays in community health and safety (Breetzke et al., 2020; Nutsford et al., 2013). These analyses have also found that increased accessibility is associated with better mental health outcomes, including anxiety, mood disorders and homicide. Increased greenspace accessibility may be indicative of increased opportunities for social cohesion (Lachowycz and Jones, 2013; Wang et al., 2021). Lack of social cohesion is known to increase the prevalence of both homicides and suicides (Nieuwbeerta et al., 2008), and can increase perceived danger of shared spaces, which may be a deterrent for accessing community greenspaces (Sreetheran and Bosch, 2014). We hypothesize that intentional deaths, most notably suicides, are lower in communities with better access to greenspaces because this access provides more avenues for social cohesion. Furthermore, our homicide results in Charlotte corroborate findings from Breetzke et al (2020); who found that homicides were lower in and around greenspaces; further suggesting that proximity to greenspace may be protective against homicidal deaths in urban areas.

Our analysis suggests that increased neighborhood greenspace accessibility in urban areas is associated with lower rates of homicide and suicide. This association was strongest in Charlotte, the city with the best greenspace accessibility, and weakest in Asheville, the city with the least greenspace accessibility. However, Asheville is in the Blue Ridge Mountains, surrounded by national forest land, which may partially explain the smaller effect greenspace accessibility had on intentional deaths in Asheville. In contrast, Charlotte, is one of the fastest growing metropolitan areas in the U.S. (Kenan Institute, 2022). Our results suggest that, as Charlotte continues to expand, city planners should ensure equitable greenspace accessibility as a low-cost community-health intervention to help increase social cohesion opportunities and reduce the prevalence of intentional deaths. For both firearm and non-firearm deaths, increasing distance to nearest greenspace was associated with a higher prevalence of intentional deaths. These findings suggest that community greenspace interventions aimed at reducing intentional deaths, regardless of manner of death (i.e., weapon type), should consider improving equitable access to existing and future greenspaces.

In all three cities, tree canopy cover was associated with lower rates of intentional deaths, though the relationship was only significant in Charlotte. Neighborhood tree canopy may act as a low cost, life-saving community health intervention. Tree canopy cover is a common metric when investigating the relationship between trees and intentional deaths and our findings corroborate past research which has found tree canopy to be protective against violent crimes and suicides (El-Mallakh et al., 2022; Gilstad-Hayden et al., 2015; Lee et al., 2023; Troy et al., 2012). New research has also documented that increased tree canopy is associated with lower direct healthcare costs in a community (Van Den Eeden et al., 2022). Our results suggest access to tree canopy is important, and in addition, the negative relationship observed with urban tree canopy cover indicates that passive health benefits (e.g., reduced psychological stress, health promotion behaviors like increased physical activity) with natural features, specifically trees, are associated with fewer intentional deaths. Tree canopy cover can provide avenues for cognitive restoration (Kaplan, 1995); which can reduce mental distress and increase mental resilience (Shepley et al., 2019; Wells and Evans, 2003). Pre-existing mental health conditions are one of the most ubiquitous factors among individuals who die by suicide and perpetrators of homicide (Flynn et al., 2016). We hypothesize that increased tree canopy cover provides opportunities for cognitive restoration, contributing to mentally resilient communities with lower rates of intentional deaths. Furthermore, increasing tree canopy cover was only significantly associated with non-firearm deaths, indicating that manner of death is an important consideration in greenspace-intentional death analyses.

Interestingly, our results found that increased quantity of greenspace, operationalized as percent public greenspace land cover, was associated with higher rates of intentional deaths. As both accessibility of greenspace and tree canopy cover were associated with reduced rates of intentional death, this contrasting finding is surprising. Our results therefore suggest that accessibility may be a more important greenspace metric to consider when conducting greenspace-health research, rather than greenspace quantity alone. Our findings support previous research which also found accessibility to be more strongly associated with poor mental health outcomes than greenspace quantity, specifically among adolescents (Markevych et al., 2017; Zach et al., 2016). Additionally, our research sheds light on how greenspace metrics can identify different relationships between greenspace and community health. As such, future research should consider multiple greenspace metrics to better understand the greenspace-health relationship. Previous research has suggested the quality, or type, of greenspace may play a key role in the greenspace-heath relationship (Wang et al., 2021; Wheeler et al., 2015). However, assessing quality is difficult. Past studies have employed the use of surveys, therefore allowing the researcher to assess self-reported high-quality greenspace and mental health outcomes (Wang et al., 2021). Other studies have qualified greenspace based on the physical characteristics of the greenspace itself, such as land cover, water features, conservation easements and species diversity and health (Wheeler et al. 2015). The findings from these studies suggest quality of greenspace may be a crucial factor in the greenspace mental health relationship. Thus, more research is needed to better understand the association between neighborhood greenspace metrics and intentional deaths.

In addition to greenspace metrics, neighborhood race was also significantly associated with intentional deaths. In Asheville and Durham, predominately Black neighborhoods were associated with higher rates of homicide, and in Charlotte, predominately Black neighborhoods were associated with higher rates of suicide. Historically, greenspace development, including urban trees, has not been equitably distributed and development of greenspaces in minority neighborhoods has often led to gentrification (Kim and Wu, 2022; Triguero-Mas et al., 2022). However, as our analysis illustrates, greenspace accessibility and tree canopy cover may act as low-cost community-health interventions. Given that intentional deaths are higher in predominantly Black neighborhoods, greenspace interventions should be targeted in minority communities, with concerted focus on greenspace development for the community. Some research cautions that active greenspace benefits the most affluent in gentrifying neighborhoods; gentrification may result in social exclusion for low-income and minority residents (Cole et al., 2019). Green gentrification, the increase in land value as a result of increased greenspace development (Anguelovski et al., 2022), has the potential to be an unintended consequence of greenspace initiatives in urban areas. Urban planning efforts must not only be aware of this consequence but must diligently invite the participation of residents in all aspects of the planning process to ensure that all residents benefit equally.

4.1. Strengths and Limitations

This analysis had several strengths. First, we had access to a high spatial resolution dataset, allowing for analysis into the accessibility of greenspaces from residential addresses, which contributes to furthering understanding of how greenspace accessibility influences the greenspace-mental health relationship. Second, our analysis moved beyond the sole use of NDVI (i.e., predominant greenspace metric in the literature) and considered three neighborhood greenspace metrics: public greenspace accessibility, public greenspace quantity, and tree canopy cover. As our results indicate, accessibility of greenspace may be more important than quantity of greenspace, and tree canopy cover is associated with benefits for community health. Third, we replicated this analysis in three distinct cities, allowing for further confirmation of the urban greenspace-health association in diverse urban settings. Finally, this study is one of the first to investigate how weapon type may influence the observed intentional death-greenspace relationship. Our results revealed that the relationship may vary based on weapon use, adding novel information to further understanding of the greenspace-health relationship.

Our analysis is also limited. First, our death data spanned 2004–19, however our greenspace metrics and covariate data were collected cross-sectionally. As such, we cannot conclude how the association between individual greenspace metrics and intentional deaths has changed over time. In addition, the composition of neighborhoods, both in terms of greenspace and socio-demographics, may have changed during the 2004–19 time period, and we cannot ensure these greenspace exposures were stable over the study period. Nonetheless, the inclusion of a long time period (2004–19) allowed us to examine the greenspace-mental health association at a fine geographic scale, geocoding residential locations, due to a larger sample size. Second, our tree canopy cover dataset is restricted to 30-meter resolution, which may not capture all of the variability in tree canopy cover in a neighborhood. Third, our analysis did not adjust for residential instability; which is commonly addressed in violent crime analyses, but not in suicide-based studies (cite?). However, we believe adjusting for racial and economic extremes provides robust results. Moreover, our analysis identified suicide and homicide locations based on residential address, and we therefore cannot conclude if these incidents occurred at a greenspace location. Similarly, greenspace metrics did not include local street geometry. Despite this limitation, this method has been conducted in other greenspace studies due to data availability (e.g., Madzia et al., 2019) and our analysis extends these methods by using exact geocoded residential addresses and a smaller geographic unit than previous work (e.g., Markevych et al., 2018). Additionally, we did not assess the quality of the greenspaces, which may influence the greenspace-health relationship. As identified in Bogar and Beyer (2016), further research considering residents’ perception of greenspace and intentional deaths is needed. Future analyses should consider neighborhood greenspace quality, in addition to greenspace quantity and accessibility.

5. Conclusions

Greenspace may be protective against intentional deaths in urban areas in the southeastern US. Increased greenspace accessibility and tree canopy cover proved to be more beneficial than increased greenspace quantity for reducing violent deaths. Our results highlight the differences in the association across greenspace metrics and researchers should be mindful of their selection in future research. In addition, our work suggests that tree canopy cover may be more important for mitigating non-firearm violent deaths; whereas improving greenspace accessibility may reduce both firearm and non-firearm intentional death prevalence. Future greenspace interventions and policy action should prioritize the accessibility of greenspace across neighborhoods. Given our results also found predominately Black neighborhoods had higher rates of homicide in Asheville and Durham, and higher rates of suicide in Charlotte, focus should be directed at improving equitable accessibility of existing and future greenspaces, especially in primarily Black neighborhoods. This study contributes new knowledge on the greenspace-health relationship and can guide future policy regarding greenspace as a low-cost community health intervention in urban areas.

Supplementary Material

MMC1

Research highlights.

  • Neighborhood greenspace accessibility is associated with fewer intentional deaths

  • Higher neighborhood tree canopy cover associated with fewer intentional deaths

  • Neighborhood greenspace accessibility association strongest for non-firearm deaths

ACKNOWLEDGEMENTS

The North Carolina Violent Death Reporting System (NC VDRS) is operated by the North Carolina Division of Public Health’s Injury and Violence Prevention Branch. NC VDRS provides detailed information on deaths resulting from violence. The findings and conclusions in this publication are those of the author(s) and do not necessarily represent the views of the North Carolina Division of Public Health’s Injury and Violence Prevention Branch. The authors have no competing interests to declare.

FUNDING

This work was supported by the Faculty Early Career Development Program (CAREER) award (grant #2044839) from the National Science Foundation and the National Institute of Environmental Health Sciences (NIEHS) award (grant # 1R15ES033817-01).

Footnotes

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

CRediT authorship contribution statement

SC Ryan: Conceptualization, Methodology, Formal Analysis, Data Curation, Writing - Original Draft, Writing - Review & Editing, Visualization.

MM Sugg: Conceptualization, Methodology, Writing - Review & Editing, Supervision, Funding Acquisition.

JD Runkle: Conceptualization, Methodology, Writing - Review & Editing, Supervision, Funding Acquisition.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. AFSP, 2022. Suicide statistics | AFSP. URL https://afsp.org/suicide-statistics/ (accessed 7.26.22). [Google Scholar]
  2. Anguelovski I, Connolly JJT, Cole H, Garcia-Lamarca M, Triguero-Mas M, Baró F, Martin N, Conesa D, Shokry G, del Pulgar CP, Ramos LA, Matheney A, Gallez E, Oscilowicz E, Máñez JL, Sarzo B, Beltrán MA, Minaya JM, 2022. Green gentrification in European and North American cities. Nat Commun 13, 3816. 10.1038/s41467-022-31572-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Anselin L, Fotheringham AS, Rogerson PA, 2008. The SAGE Handbook of Spatial Analysis. SAGE. [Google Scholar]
  4. Beard JH, Morrison CN, Jacoby SF, Dong B, Smith R, Sims CA, Wiebe DJ, 2017. Quantifying Disparities in Urban Firearm Violence by Race and Place in Philadelphia, Pennsylvania: A Cartographic Study. Am J Public Health 107, 371–373. 10.2105/AJPH.2016.303620 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Beyer KMM, Kaltenbach A, Szabo A, Bogar S, Nieto FJ, Malecki KM, 2014. Exposure to neighborhood green space and mental health: evidence from the survey of the health of Wisconsin. Int J Environ Res Public Health 11, 3453–3472. 10.3390/ijerph110303453 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bogar S, Beyer KM, 2016. Green Space, Violence, and Crime: A Systematic Review. Trauma, Violence, & Abuse 17, 160–171. 10.1177/1524838015576412 [DOI] [PubMed] [Google Scholar]
  7. Breetzke G, Pearson A, Tao S, Zhang R, 2020. Greenspace and Gun Violence in Detroit, USA. International Journal of Criminal Justice Sciences 15, 248–265. 10.5281/zenodo.3865608 [DOI] [Google Scholar]
  8. Browning MHEM, Rigolon A, 2018. Do Income, Race and Ethnicity, and Sprawl Influence the Greenspace-Human Health Link in City-Level Analyses? Findings from 496 Cities in the United States. Int J Environ Res Public Health 15, E1541. 10.3390/ijerph15071541 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Browning MHEM, Rigolon A, Ogletree S, Wang R, Klompmaker JO, Bailey C, Gagnon R, James P, 2022. Curation of new green space indicator for the U.S.: Accessible & recreational park cover (PAD-US-AR). 10.32942/osf.io/pydqe [DOI]
  10. Cole HVS, Triguero-Mas M, Connolly JJT, Anguelovski I, 2019. Determining the health benefits of green space: Does gentrification matter? Health Place 57, 1–11. 10.1016/j.healthplace.2019.02.001 [DOI] [PubMed] [Google Scholar]
  11. Collins R, Spake R, Brown KA, Ogutu B, Smith D, Eigenbrod F, 2020. A systematic map of research exploring the effect of greenspace on mental health [WWW Document]. Landscape and Urban Planning. 10.1016/j.landurbplan.2020.103823 [DOI] [Google Scholar]
  12. Craney TA, Surles JG, 2002. Model-Dependent Variance Inflation Factor Cutoff Values. Quality Engineering 14, 391–403. 10.1081/QEN-120001878 [DOI] [Google Scholar]
  13. Duncan DT, Kawachi I (Eds.), 2018. Neighborhoods and Health, 2nd ed. Oxford University Press, New York. 10.1093/oso/9780190843496.001.0001 [DOI] [Google Scholar]
  14. Ekkel ED, de Vries S, 2017. Nearby green space and human health: Evaluating accessibility metrics. Landscape and Urban Planning 157, 214–220. 10.1016/j.landurbplan.2016.06.008 [DOI] [Google Scholar]
  15. El-Mallakh TVL, Hedges S, Rai J, Bhatnagar A, Moyer S, El-Mallakh R, 2022. Suicide and Homicide More Common with Limited Urban Tree Canopy Cover. CATE 14. 10.15365/cate.2022.140204 [DOI] [Google Scholar]
  16. ESRI, 2022. ArcGIS Pro 2.9.0 [WWW Document]. URL https://www.esri.com/en-us/arcgis/products/arcgis-pro/overview (accessed 3.21.22).
  17. Flynn S, Gask L, Appleby L, Shaw J, 2016. Homicide–suicide and the role of mental disorder: a national consecutive case series. Soc Psychiatry Psychiatr Epidemiol 51, 877–884. 10.1007/s00127-016-1209-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gilstad-Hayden K, Wallace LR, Carroll-Scott A, Meyer SR, Barbo S, Murphy-Dunning C, Ickovics JR, 2015. Research note: Greater tree canopy cover is associated with lower rates of both violent and property crime in New Haven, CT. Landscape and Urban Planning 143, 248–253. 10.1016/j.landurbplan.2015.08.005 [DOI] [Google Scholar]
  19. Helbich M, de Beurs D, Kwan M-P, O’Connor RC, Groenewegen PP, 2018. Natural environments and suicide mortality in the Netherlands: a cross-sectional, ecological study. The Lancet Planetary Health 2, e134–e139. 10.1016/S2542-5196(18)30033-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Houlden V, Porto de Albuquerque J, Weich S, Jarvis S, 2021. Does nature make us happier? A spatial error model of greenspace types and mental wellbeing. Environment and Planning B: Urban Analytics and City Science 48, 655–670. 10.1177/2399808319887395 [DOI] [Google Scholar]
  21. Houlden V, Porto de Albuquerque J, Weich S, Jarvis S, 2019. A spatial analysis of proximate greenspace and mental wellbeing in London. Applied Geography 109, 102036. 10.1016/j.apgeog.2019.102036 [DOI] [Google Scholar]
  22. James G, Witten D, Hastie T, Tibshirani R, 2013. An introduction to statistical learning with applications in R. Springer Science and Business Media, New York. [Google Scholar]
  23. Jiang W, Stickley A, Ueda M, 2021. Green space and suicide mortality in Japan: An ecological study. Social Science & Medicine 282, 114137. 10.1016/j.socscimed.2021.114137 [DOI] [PubMed] [Google Scholar]
  24. Kaplan S, 1995. The restorative benefits of nature: Toward an integrative framework. Journal of Environmental Psychology, Green Psychology 15, 169–182. 10.1016/0272-4944(95)90001-2 [DOI] [Google Scholar]
  25. Kenan Institute, 2022. The American Growth Project. UNC Kenan-Flagler Business School. [Google Scholar]
  26. Kim SK, Wu L, 2022. Do the characteristics of new green space contribute to gentrification? Urban Studies 59, 360–380. 10.1177/0042098021989951 [DOI] [Google Scholar]
  27. Krieger N, Waterman PD, Spasojevic J, Li W, Maduro G, Van Wye G, 2016. Public Health Monitoring of Privilege and Deprivation With the Index of Concentration at the Extremes. Am J Public Health 106, 256–263. 10.2105/AJPH.2015.302955 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kuo FE, Sullivan WC, 2001. Environment and Crime in the Inner City: Does Vegetation Reduce Crime? Environment and Behavior 33, 343–367. 10.1177/0013916501333002 [DOI] [Google Scholar]
  29. Lachowycz K, Jones A, 2013. Towards a better understanding of the relationship between greenspace and health: Development of a theoretical framework. Landscape and Urban Planning 118, 62–69. 10.1016/j.landurbplan.2012.10.012 [DOI] [Google Scholar]
  30. Lee S, Lee RJ, Scherr S, 2023. How tree canopy cover can reduce urban suicide attempts: A geospatial analysis of the moderating role of area deprivation. Landscape and Urban Planning 230, 104606. 10.1016/j.landurbplan.2022.104606 [DOI] [Google Scholar]
  31. Legendre P, 1993. Spatial Autocorrelation: Trouble or New Paradigm? Ecology 74, 1659–1673. 10.2307/1939924 [DOI] [Google Scholar]
  32. Madzia J, Ryan P, Yolton K, Percy Z, Newman N, LeMasters G, Brokamp C, 2019. Residential Greenspace Association with Childhood Behavioral Outcomes. The Journal of Pediatrics 207, 233–240. 10.1016/j.jpeds.2018.10.061 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Markevych I, Schoierer J, Hartig T, Chudnovsky A, Hystad P, Dzhambov AM, de Vries S, Triguero-Mas M, Brauer M, Nieuwenhuijsen MJ, Lupp G, Richardson EA, Astell-Burt T, Dimitrova D, Feng X, Sadeh M, Standl M, Heinrich J, Fuertes E, 2017. Exploring pathways linking greenspace to health: Theoretical and methodological guidance. Environ Res 158, 301–317. 10.1016/j.envres.2017.06.028 [DOI] [PubMed] [Google Scholar]
  34. Markevych I, Tesch F, Datzmann T, Romanos M, Schmitt J, Heinrich J, 2018. Outdoor air pollution, greenspace, and incidence of ADHD: A semi-individual study. Sci Total Environ 642, 1362–1368. 10.1016/j.scitotenv.2018.06.167 [DOI] [PubMed] [Google Scholar]
  35. Mears M, Brindley P, 2019. Measuring Urban Greenspace Distribution Equity: The Importance of Appropriate Methodological Approaches. ISPRS International Journal of Geo-Information 8, 286. 10.3390/ijgi8060286 [DOI] [Google Scholar]
  36. NCDHHS, 2022. N.C. DPH: IVP Branch: Violent Death Data [WWW Document]. Chronic Disease and Injury Section. URL https://injuryfreenc.dph.ncdhhs.gov/DataSurveillance/ViolentDeathData.htm (accessed 7.8.22). [Google Scholar]
  37. Nieuwbeerta P, McCall PL, Elffers H, Wittebrood K, 2008. Neighborhood Characteristics and Individual Homicide Risks: Effects of Social Cohesion, Confidence in the Police, and Socioeconomic Disadvantage. Homicide Studies 12, 90–116. 10.1177/1088767907310913 [DOI] [Google Scholar]
  38. Nutsford D, Pearson A, Kingham S, 2013. An ecological study investigating the association between access to urban green space and mental health. Public health 127. 10.1016/j.puhe.2013.08.016 [DOI] [PubMed] [Google Scholar]
  39. Parent MC, Hammer JH, Bradstreet TC, Schwartz EN, Jobe T, 2018. Men’s Mental Health Help-Seeking Behaviors: An Intersectional Analysis. Am J Mens Health 12, 64–73. 10.1177/1557988315625776 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. RStudio Team, 2022. RStudio: Integrated Development Environment for R.
  41. Sadatsafavi H, Sachs NA, Shepley MM, Kondo MC, Barankevich RA, 2022. Vacant lot remediation and firearm violence – A meta-analysis and benefit-to-cost evaluation. Landscape and Urban Planning 218, 104281. 10.1016/j.landurbplan.2021.104281 [DOI] [Google Scholar]
  42. Sarkar C, Webster C, Gallacher J, 2018. Residential greenness and prevalence of major depressive disorders: a cross-sectional, observational, associational study of 94 879 adult UK Biobank participants. Lancet Planet Health 2, e162–e173. 10.1016/S2542-5196(18)30051-2 [DOI] [PubMed] [Google Scholar]
  43. Shen Y-S, Lung S-CC, 2018. Identifying critical green structure characteristics for reducing the suicide rate. Urban Forestry & Urban Greening 34, 147–153. 10.1016/j.ufug.2018.06.005 [DOI] [Google Scholar]
  44. Shepley M, Sachs N, Sadatsafavi H, Fournier C, Peditto K, 2019. The Impact of Green Space on Violent Crime in Urban Environments: An Evidence Synthesis. International Journal of Environmental Research and Public Health 16, 5119. 10.3390/ijerph16245119 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Sreetheran M, Bosch CCK van den, 2014. A socio-ecological exploration of fear of crime in urban green spaces – A systematic review. Urban Forestry &amp; Urban Greening 13, 1. [Google Scholar]
  46. The Trust for Public Land, 2021. ParkServe Data Downloads [WWW Document]. The Trust for Public Land. URL https://www.tpl.org/parkserve/downloads (accessed 11.29.21). [Google Scholar]
  47. Triguero-Mas M, Anguelovski I, Connolly JJT, Martin N, Matheney A, Cole HVS, Pérez-Del-Pulgar C, García-Lamarca M, Shokry G, Argüelles L, Conesa D, Gallez E, Sarzo B, Beltrán MA, López Máñez J, Martínez-Minaya J, Oscilowicz E, Arcaya MC, Baró F, 2022. Exploring green gentrification in 28 global North cities: the role of urban parks and other types of greenspaces. Environ. Res. Lett. 17, 104035. 10.1088/1748-9326/ac9325 [DOI] [Google Scholar]
  48. Troy A, Morgan Grove J, O’Neil-Dunne J, 2012. The relationship between tree canopy and crime rates across an urban–rural gradient in the greater Baltimore region. Landscape and Urban Planning 106, 262–270. 10.1016/j.landurbplan.2012.03.010 [DOI] [Google Scholar]
  49. UNODC, 2019. Homicide mechanisms and enablers. United Nations Office on Drugs and Crimes. [Google Scholar]
  50. US Census, 2018. American Community Survey (ACS) [WWW Document]. Census.gov. URL https://www.census.gov/programs-surveys/acs (accessed 5.10.22).
  51. U.S. Census Bureau, 2022. U.S. Census Bureau QuickFacts: Asheville city, North Carolina [WWW Document]. URL https://www.census.gov/quickfacts/ashevillecitynorthcarolina (accessed 5.16.23).
  52. U.S. Geological Survey (USGS) Gap Analysis Project (GAP), 2020. Protected Areas Database of the United States (PAD-US) 2.0. 10.5066/P955KPLE [DOI]
  53. USFS, 2019. USFS Tree Canopy Cover Datasets.
  54. Van Den Eeden SK, Browning HEM, Becker M, Shan DA, Alexeeff J, Thomas Ray SE, Quesenberry G, Kuo CP, M., 2022. Association between residential green cover and direct healthcare costs in Northern California: An individual level analysis of 5 million persons. Environment International 163, 107174. 10.1016/j.envint.2022.107174 [DOI] [PubMed] [Google Scholar]
  55. Wang R, Feng Z, Pearce J, Liu Y, Dong G, 2021. Are greenspace quantity and quality associated with mental health through different mechanisms in Guangzhou, China: A comparison study using street view data. Environmental Pollution 290, 117976. 10.1016/j.envpol.2021.117976 [DOI] [PubMed] [Google Scholar]
  56. Wells NM, Evans GW, 2003. Nearby Nature: A Buffer of Life Stress among Rural Children. Environment and Behavior 35, 311–330. 10.1177/0013916503035003001 [DOI] [Google Scholar]
  57. Wheeler BW, Lovell R, Higgins SL, White MP, Alcock I, Osborne NJ, Husk K, Sabel CE, Depledge MH, 2015. Beyond greenspace: an ecological study of population general health and indicators of natural environment type and quality. International Journal of Health Geographics 14, 17. 10.1186/s12942-015-0009-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Wilkins NJ, Zhang X, Mack KA, Clapperton AJ, Macpherson A, Sleet D, Kresnow-Sedacca M, Ballesteros MF, Newton D, Murdoch J, Mackay JM, Berecki-Gisolf J, Marr A, Armstead T, McClure R, 2019. Societal determinants of violent death: The extent to which social, economic, and structural characteristics explain differences in violence across Australia, Canada, and the United States. SSM - Population Health 8, 100431. 10.1016/j.ssmph.2019.100431 [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Zach A, Meyer N, Hendrowarsito L, Kolb S, Bolte G, Nennstiel-Ratzel U, Stilianakis NI, Herr C, 2016. Association of sociodemographic and environmental factors with the mental health status among preschool children—Results from a cross-sectional study in Bavaria, Germany. International Journal of Hygiene and Environmental Health 219, 458–467. 10.1016/j.ijheh.2016.04.012 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

MMC1
NIHMS1914302-supplement-MMC1.docx (39.2KB, docx)

RESOURCES