Beyond Social Disadvantage: Advancing an Environmental Justice Framework to Address Child Maltreatment Risk

Beyond Social Disadvantage: Advancing an Environmental Justice Framework to Address Child Maltreatment Risk | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Beyond Social Disadvantage: Advancing an Environmental Justice Framework to Address Child Maltreatment Risk Gia Barboza-Salerno, Sharefa Duhaney, Balaji Ramesh, Charis Stanek This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6521185/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 09 Apr, 2026 Read the published version in International Journal on Child Maltreatment: Research, Policy and Practice → Version 1 posted You are reading this latest preprint version Abstract Child abuse and neglect (CAN) represent a significant global public health challenge influenced by socioeconomic disadvantages and the built environment. While existing research has looked into the social determinants of CAN, fewer studies have focused on how environmental factors interact with social vulnerabilities to affect risk levels. This study aims to fill that gap by utilizing an environmental justice framework and Self-Organizing Maps (SOMs) to categorize neighborhoods in Los Angeles based on social, environmental, and health-related characteristics. We examined physical abuse (CPA) and child neglect (CN) from 2020 to 2023 at the census block group level. Fifteen georeferenced indicators—such as poverty, tree equity, park access, heat exposure, and mental health—were used as input features in a Self-Organizing Map (SOM) to identify clusters of neighborhoods with similar socio-environmental profiles. Negative binomial regression was used to predict CN and CPA rates within clusters. Seven clusters describe socio-environmental neighborhood profiles in Los Angeles. The most disadvantaged cluster was defined by high poverty rates, limited green space equity, and poor mental health, with CPA and CN rates more than double those of the most advantaged cluster. Risk levels were significantly higher in areas with intersecting social and environmental challenges. These findings highlight that structural inequities, including restricted access to green infrastructure, increase CAN risk. Our results suggest the need for targeted investments in parks, trees, and other features of the built environment in underserved neighborhoods as part of a comprehensive, place-based approach supporting healthy environments for children. child abuse and neglect environmental justice built environment neighborhood risk tree equity green space public health self-organizing maps structural inequity Los Angeles Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Introduction Child abuse and neglect are significant public health concerns with lasting consequences for both children and communities. In 2021, U.S. Child Protective Services (CPS) investigated over 3 million cases of suspected maltreatment, with an estimated 600,000 children confirmed as victims of abuse or neglect ( New Child Maltreatment Report Finds Child Abuse and Neglect Decreased to a Five-Year Low , n.d.; U.S. Department of Health & Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau, 2024). Neglect accounted for the majority of these cases, with 70% of substantiated victims under the age of ten (Foundation, 2021). In Los Angeles County, the focus of this study, African American/Black children experience child abuse and neglect reporting rates that are 3.4 times higher than their White counterparts ( Reports of Child Abuse and Neglect, by Race/Ethnicity , n.d.). While systemic inequities, including structural racism, contribute to these disparities, few studies explicitly examine how neighborhood socio-environmental conditions shape child maltreatment risk (Littleton et al., 2024). Even fewer apply frameworks that analyze how structural racism operates through socio-environmental characteristics (Foundation, 2021; Riley et al., 2024). The social determinants of health within neighborhoods significantly influence child welfare outcomes through interconnected economic, social, and environmental pathways. Previous research has established a correlation between child abuse and neglect and various social and economic risk factors, including elevated poverty and unemployment rates (Barboza-Salerno, 2024), housing instability (Barboza-Salerno, 2020b; Littleton et al., 2024), restricted access to essential services (Maguire-Jack & Font, 2017), and a high concentration of alcohol and drug outlets (Freisthler, 2004; Freisthler et al., 2022). Although social and economic conditions are pivotal, the built environment can either exacerbate or alleviate these risks, depending on the broader social context. Nevertheless, existing research has predominantly overlooked the interplay between the built environment, such as access to green spaces, walkability, and recreational opportunities, and factors associated with Child Protective Services (CPA) and Child Neglect (CN) risk. A notable exception is a recent study that identified an increase in the risk of child welfare involvement related to tree canopy cover and green space, even when controlling for neighborhood social conditions, including area-level deprivation (He et al., 2024). However, the built environment encompasses a range of distinct features—including extreme heat (Evans et al., 2025; Le, 2025), morphology (e.g., street layout) (Barboza-Salerno & Meshelemiah, 2023; Chen et al., 2021), and equitable access to trees and parks—that collectively influence behavioral patterns, movement, and social interactions pertinent to CPA and CN risk. Noah (2015, p. 457) characterizes the failure to integrate environmental and social exposures as “spatial polygamy,” wherein isolated analyses disregard the cumulative effects of overlapping risk factors. The purpose of this study is to examine whether neighborhoods with similar social, economic, and physical environmental characteristics exhibit comparable patterns of child physical abuse and neglect risk in Los Angeles, California. To achieve this, we used multiple indicators of the social, economic, and physical environment to classify areas into distinct typologies using Self-Organizing Maps (SOMs). We then mapped the typologies to geographic space to compare small-area rates of CPA and CN within each cluster. Unlike traditional distance-based clustering models, which rely on predefined assumptions about variable relationships, SOMs employ an unsupervised machine learning algorithm to detect nonlinear patterns and emergent zonal structures. This study develops a new framework for understanding the geographic and contextual factors influencing child victimization risk by incorporating less-explored socio-environmental elements, such as equitable park access, urban heat islands, and intersection density, alongside traditional socioeconomic measures. Association Between Neighborhood Social, Physical, and Environmental Characteristics and Child Abuse and Neglect A substantial body of research has documented the individual (Azar & Weinzierl, 2005; Cause, 2020; Maguire-Jack & Font, 2017) and contextual factors (Barboza-Salerno, 2023; Elise Barboza-Salerno, 2024; Font & Maguire-Jack, 2015; Freisthler, 2004; Freisthler et al., 2003; Littleton et al., 2024; Maguire-Jack et al., 2022) associated with CPA and CN. At the individual level, the risk is closely linked to family characteristics, including a child’s age and gender, the caregiver's mental health, material deprivation, substance misuse, and socioeconomic status. Parenting stress, in particular, has been strongly associated with child maltreatment, often mediated by child behavioral problems and poor socioemotional adjustment (Barboza & Schiamberg, 2021; Barboza-Salerno, 2020a). At the contextual level, similar social factors, such as poverty, material deprivation, and residential instability, also play a critical role in shaping maltreatment risk. Moreover, evidence suggests a graded, dose-response relationship between social vulnerability and child maltreatment, with greater exposure to adverse social conditions correlating with increased risk for violence or abuse (Barboza et al., 2022; Barboza-Salerno, 2023; Bywaters et al., 2016). Research suggests that the built environment, defined as “the human-made space where people live, work, and recreate on a day-to-day basis (Roof & Oleru, 2008),” moderates the socioeconomic structure of neighborhoods with important consequences for child well-being (Morton et al., 2014). The built environment encompasses buildings (residential, commercial, and industrial), infrastructure (roads, bridges, water systems, and energy grids), public spaces (parks, plazas, and playgrounds), and landscapes shaped by urban and rural development. A qualitative study examining caregivers’ views on the built environment's influence on child maltreatment revealed that abandoned buildings, insufficient green space, and poorly maintained infrastructure correlate with a higher risk of child maltreatment (Haas et al., 2018). Abandoned buildings and deteriorating infrastructure can contribute to neighborhood disorder, increasing parental stress by fostering unsafe conditions, limiting access to community resources, and reducing opportunities for social support (Guterman et al., 2009; Maguire-Jack & Showalter, 2016). In neighborhoods with limited green spaces, children face an increased risk of poor health outcomes, such as obesity and respiratory issues, due to the role of reduced vegetation in creating urban heat islands (Moudon et al., 2006). Research has shown that proximity to parks is associated with reduced child welfare involvement, lower parenting stress, increased social interactions, and greater physical activity (Frumkin et al., 2017; Kuo & Sullivan, 2001a, 2001b). Densely populated areas with high rental occupancy and shared living spaces can exacerbate maltreatment risks due to overcrowding, unsafe conditions, and limited supervision. Studies have highlighted how urban sprawl—characterized by imbalanced housing and employment opportunities—may heighten parenting stress and negatively impact child outcomes. Access to parks and open spaces fosters social interaction and reduces stress (Fan et al., 2011; Kendrick, 2015; Morris et al., 2017), while overcrowded and unsafe living conditions in high-density areas contribute to increased maltreatment risk (Finno-Velasquez et al., 2017). Furthermore, physical isolation—often resulting from neglected community spaces or barriers to social engagement—exacerbates vulnerabilities by limiting access to essential resources and social support networks (Chung et al., 2022; Ellis et al., 2020). The benefits of living near green spaces are not uniform but vary across different socioeconomic contexts. Therefore, the effect modification of green space on the association between parenting stress and child well-being depends on the broader social conditions that shape parent-child interactions. For example, a study of 1,468 mothers in Lithuania found that greater distances from parks were associated with worse mental health outcomes but only among parents with lower educational attainment. While green spaces offer potential protective effects, other aspects of the built environment may also exacerbate child physical abuse and neglect risk. Morton et al. (2014) found that areas with easier access to substance abuse services had lower rates of neglect, even after controlling for neighborhood demographics and socioeconomic structure. Further, the presence of substance abuse service facilities moderated the impact of alcohol outlet density on child maltreatment rates. Despite these associations, research on child welfare determinants has been fragmented, often overlooking the complex, non-linear interactions between social, economic, and environmental factors that may play a causal role in child victimization (Balseviciene et al., 2014), with only a handful of studies focusing on the broader environmental characteristics contributing to increased risk (Haas et al., 2018). A theoretical and conceptual model Social disorganization theory has traditionally guided sociological and social work perspectives on child maltreatment risk (Green, 2022). However, recent research indicates that the physical characteristics of the built environment also play a significant role, expanding our understanding beyond just socioeconomic and social psychological factors. The focus on socioeconomic factors related to child well-being, particularly poverty, has created a lack of conceptual clarity about how the built environment may contribute to both child poverty and neighborhood disadvantage. According to SDT, neighborhood social disorder disrupts social cohesion and weakens informal social control, reducing collective efficacy and ultimately contributing to higher levels of crime, violence, and abuse. Although SDT’s fundamental assertion that neighborhood environments influence child outcomes is well-supported, emerging research indicates that these effects are primarily driven by systemic forces that shape physical and social well-being (Littleton et al., 2024). While SDT focuses on the consequences of neighborhood disorder, it largely overlooks these systemic factors. Decades of redlining and discriminatory urban planning practices have resulted in neighborhoods where social vulnerability overlaps with environmental risk, compounding stressors that increase the likelihood of child maltreatment (Littleton et al., 2024). Community disinvestment erodes essential social support and resources, particularly in impoverished urban areas. Recognizing these structural forces shifts the focus from neighborhood collective efficacy to systemic harm created and sustained by de jure segregation. On this basis, Ruth Gilmore’s organized abandonment theory offers a more compelling lens for understanding how neighborhoods contribute to CPA and CN. According to organized abandonment theory, broader patterns of intentional systemic disinvestment create conditions where children face heightened vulnerability and exposure to violence. Unlike SDT, which places the responsibility for resolving neighborhood social and physical ‘disorganization’ (renaming the word to organization does not change the theory’s assumptions) on residents, organized abandonment theory highlights systemic inequities, such as disparities in green space access, poor housing conditions, and environmental burdens, that require alternative law and policy responses. Using organized abandonment as a lens to view the mechanisms linking social and physical conditions to CPA and CN, we introduce the conceptual model shown in Fig. 1 . In our model, organized abandonment primarily operates through systematic disinvestment, creating neighborhoods vulnerable to social, environmental, and health burdens. In turn, socially vulnerable areas have unequal access to high-quality public spaces due to limited walkability and urban decay. Communities facing socioeconomic disadvantages often experience higher environmental burdens, such as exposure to pollution and a lack of green spaces, which compound risks to children's well-being. Neighborhood vulnerability and green space morphology structure interactions between parents and children in specific contexts. Our conceptual model is theoretically grounded in Bronfenbrenner’s ecological systems theory and Garbarino’s human ecology model, which frames child maltreatment as the result of interacting influences across multiple levels, from individual interactions (microsystem) to broader societal structures (macrosystem). Within this framework, natural environment features (e.g., access to green spaces) are critical components that shape child outcomes through public spaces where children and their parents interact. Equitable access to green spaces, walkable streets, and essential amenities is crucial for positive social interactions. Several potential mechanisms explain how these characteristics may heighten CAN risk. First, green spaces, walkability, and community resources reduce parental stress and foster social connections, thereby minimizing the risks of child maltreatment (Balseviciene et al., 2014). Neighborhoods with abundant green spaces may provide safe environments for children to play, socialize, and engage in physical activity, promoting healthy development and reducing CAN risk (Fan et al., 2011). In contrast, neighborhoods with limited or unsafe parks often confine children indoors for extended periods, which increases parental stress, supervision challenges, and risks of both CPA and CN. Additionally, deteriorating or unsafe public spaces expose children to physical hazards, further heightening maltreatment risks (Guterman et al., 2009). The conceptual model informs our methodological approach by identifying key variables (e.g., park equity, intersection density, physical inactivity, and poverty) that serve as input features for training our machine learning model. Current Study This study employs SOMs, an unsupervised machine-learning technique that identifies nonlinear relationships between socioenvironmental and geographic factors to understand how underlying exposure patterns interact in complex, dynamic systems (Basara & Yuan, 2008). By clustering areas with similar characteristics, SOMs reveal hidden patterns in high-dimensional data, providing a more nuanced understanding of how these areas collectively influence child maltreatment risk. Using SOMs, we construct a typology that links spatial patterns of child maltreatment to specific environmental features. By examining the relative risks of CPA and CN across these clusters, we seek to answer two research questions: (a) What neighborhood typologies arise from grouping the built, natural, and social environments in Los Angeles? Furthermore, (b) Which neighborhood typologies have higher rates of CPA and CN? This approach provides insights into socio-environmental factors that shape CPA and CN risk, highlighting the need for prevention strategies tailored to neighborhood-specific needs. Materials and Methods Study Population and Design This study uses incident data from the Los Angeles Police Department on CN and CPA occurring between January 1, 2020, and December 31, 2023. Under California law, child neglect (Penal Code § 270 PC) is when a parent or legal guardian willfully (and without lawful excuse) fails to provide necessities such as clothing, food, medicine, and shelter. Child abuse (Penal Code § 273d PC) is defined as the willful infliction of cruel or inhuman corporal punishment or causing a traumatic injury (e.g., slapping, punching, or hitting a child). To calculate CPA and CN rates, we obtained 5-year population estimates for individuals aged 0 to 17 within each census block group from the American Community Survey (ACS) for 2018–2023. We merged environmental data on tree and park equity and access from American Forests via the Tree Equity Score, the California Office of Environmental Health Hazard Assessment, and the Trust for Public Land’s ParkServe platform. Additionally, physical and mental health indicators were drawn from the Centers for Disease Control and Prevention’s (CDC) PLACES dataset, derived from the Behavioral Risk Factor Surveillance System (BRFSS), to capture neighborhood-level health factors that may be linked to CPA and CN outcomes. Dependency information, including the ratio of adults over 65 to children under 18, linguistic isolation, poverty, race, and unemployment, was provided by the ACS 5-year estimates (see Table 1 ). Table 1 Indicators Dataset Metric Type Source Timeframe Child Physical Abuse and Neglect Health – primary outcome variable City of Los Angeles Open Data Portal – Crime Incidents from 2000 to present (Crime Data from 2020 to Present | Los Angeles - Open Data Portal) 2023 Tree Equity Score (TES) Environmental Burden American Forests (Tree Equity - American Forests) 2023 Heat Extremity Environmental Burden Trust for Public Land’s ParkServe Database (ParkServe - Trust for Public Land) 2023 Park Equity Score (PES) Environmental Burden 2023 Green Space along road network (percent green space) Environmental Burden EnviroAtlas (EnviroAtlas Data | US EPA) 2018 Tree Cover along road network (percent cover) Environmental Burden Physical Inactivity Health Burden U.S. Centers for Disease Control and Prevention (CDC) 500 Cities Project, renamed PLACES as of late 2020 (Centers for Disease Control and Prevention et al., 2023) 2023 Poor Mental Health Health Burden 2023 Age Dependency Social Burden U.S. American Community Survey (ACS) (U.S. Census Bureau, 2023) 2019–2023 Unemployment (%) Social Burden 2019–2023 Poverty (%) Social Burden 2019–2023 Linguistic Isolation (%) Social Burden 2019–2023 Non-White Residents (%) Social Burden 2019–2023 Percent of Children (%) Social Burden 2019–2023 Environmental burden. Tree- and Park Equity . The tree equity score (TES; Fig. 2 A) developed by American Forests is used here to evaluate the impact of the distribution of urban tree canopy on CPA and CN risk. The TES is constructed by considering tree canopy goals for each area and identifying priority areas associated with the greatest need for tree coverage. Neighborhood tree canopy goals are calculated based on natural biome baselines (e.g., forest: 40%, grassland: 20%, desert: 15%) adjusted by building density, which limits spaces where trees can be planted. Tree canopy cover is obtained from a pre-aggregated high-resolution tree canopy dataset provided by Google Environmental Insights Explorer. The percentage of tree canopy cover is calculated as follows: $$\:Canopy\:Cover\:\left(\%\right)=\raisebox{1ex}{${A}_{T}$}\!\left/\:\!\raisebox{-1ex}{${A}_{L},$}\right.\times\:100$$ , where: \(\:{A}_{L}\) is the land area of the block group, not including the water area, and \(\:{A}_{T}\) is the tree canopy cover in the area. Building density adjustments utilize a trend line between density and canopy for each biome, with forests and Mediterranean areas receiving the most significant adjustments. The canopy gap (GAP) measures the difference between the existing tree canopy and the adjusted goal. The TES is calculated by multiplying the gap score with an index, E , for prioritizing neighborhoods with the greatest need for tree planting: $$\:TES=100\left(1-{GAP}_{score}\times\:E\right)$$ . Scores range from 0 to 100, with higher scores reflecting greater equity. Prior research defines Moderate Tree Equity as 70–90% of tree canopy cover; High Tree Equity is more than 90%. Park need was categorized based on the following thresholds: Very High Park Need: Park need greater than 3.5; High Park Need: Park need between 3 and 3.5; Moderate Park Need: Park need between 2 and 3; Low Park Need: Park need less than 2. Heat extremity TPL’s urban heat severity is based on the highest extremities of summer surface temperatures averaged by block group. It is defined as the average land surface temperature of the block group in degrees Fahrenheit above or below the city's average land surface temperature (TPL Urban Heat Islands dataset 2023). Green Space within twenty-five meters of the road centerline Green space is defined as areas containing trees and forests, shrubs, grasses, herbaceous plants, woody wetlands, and emergent wetlands, with water bodies along walkable roads provided by the EnviroAtlas (2020). The EnviroAtlas dataset measures green space within twenty-five meters of each road centerline and provides a percentage of the green space relative to the total area between street intersections. Tree Cover within an 8.5-meter strip starting from the road edge The percentage of tree cover containing trees, forests, and woody wetlands along walkable roads provided by the EnviroAtlas (2020). Tree cover is calculated within an 8.5-meter strip starting from the road edge (Fig. 2 B). Low tree canopy cover was defined as less than 70% of the total tree canopy cover. Park access. Walking access to parks was assessed using the Natural Lands Trust tool called ParkServe®, which identifies the percent of a community within a 10-minute walk (1/4 mile buffer of a park. Health Vulnerability . Physical inactivity represents the percentage of adults (aged eighteen and older) in a census tract who report not participating in any physical activities or exercises, such as running, walking, or gardening, outside their regular job duties over the past month. Poor mental health was measured as the percentage of adults in a census block group (CBG) who reported experiencing poor mental health for 14 or more days within the past 30 days. Poor mental health includes feelings of stress, depression, or emotional challenges that affect daily functioning. These estimates are based on 2020 BRFSS data and modeled for small geographic areas. Socioeconomic Vulnerability . Age dependency is calculated as the ratio of non-working populations, including children (0–17) and seniors (65+), to the working-age population, reflecting societal support demands. Unemployment is defined as the percentage of people in the CBG who are not working, are actively seeking employment, and are available to work. Poverty is measured as the percentage of individuals living below 200% of the federal poverty line. Linguistic isolation is the proportion of households where no individual aged fourteen or older speaks English fluently. Race and ethnicity are captured by the percentage of populations identifying as non-Latine White, including persons who self-identify as Black, American Indian, Alaska Native, Asian, Native Hawaiian, Pacific Islander, and Latine. Statistical Analysis Aligning with the conceptual model presented in Fig. 1 , this study employs SOMs to identify patterns in environmental, social, and health burdens, such as Tree Equity Score, physical inactivity, poor mental health, and access to green space, that are associated with the risk of CPA and CN. A SOM is an artificial neural network (ANN)-based classifier that utilizes an unsupervised machine learning algorithm to cluster neighborhoods with similar characteristics, revealing hidden patterns that traditional statistical methods may overlook (Basara & Yuan, 2008; Kohonen, 1990). To build the SOM, a predefined surface is initialized where each neuron (or unit) is assigned a weight vector corresponding to the dimensionality of the input features. During training, the Best Matching Unit (BMU)—the neuron with the smallest distance to an input vector—is identified, and both the BMU and its neighboring neurons adjust their weight vectors toward the input data (Augustijn & Zurita-Milla, 2013; Basara & Yuan, 2008). While the SOM does not directly incorporate geographic coordinates, it preserves spatial coherence by grouping feature-similar and spatially proximate neighborhoods into adjacent regions of the map. This ensures the final SOM clusters retain meaningful geographic and contextual relationships when projected back into physical space (Andrienko et al., 2010). As a result, spatially closer neighborhoods with similar characteristics are more likely to be grouped, forming geographically meaningful clusters when visualized on a map. For this analysis, fifteen variables were normalized using Z-scores, and multicollinearity was assessed before training the SOM. To train the SOM, we used a 20 × 10 hexagonal grid, with each node representing a distinct region of the feature space. The training consisted of an initial exploratory phase followed by an extended iteration phase (rlen = 1000) to refine node weight vectors and minimize the mean distance between node weights and their assigned observations. As training progressed, data points were mapped to their BMUs, and nearby neurons adjusted their weights, preserving topological relationships. Over time, the SOM converged to a structured 2D representation of the original high-dimensional feature space. We used k -means clustering to determine the optimal number of clusters based on the within-cluster sum of squares (WSS). These clusters were visualized with boundaries to highlight their spatial structure. We examined the loss function plot to evaluate convergence, and changes in node distances were visualized to identify potential cluster boundaries. Additionally, component plane plots were used to interpret the SOM by mapping key variables, such as park priority, unemployment rate, and tree cover, onto the SOM nodes, helping to reveal underlying patterns in environmental and social risk factors associated with CPA and CN. After training, the SOM nodal assignments were merged with the incident-level data to categorize each incident by its SOM cluster. We then used a negative binomial regression model to assess the relationship between SOM clusters and rates of CPA and CN. The dependent variables in the model were the number of CPA and CN cases recorded within each CBG. Given that the outcome variables are counts and the counts were overdispersed within CBGs, a negative binomial model was chosen to better account for the relatively higher variance-to-mean ratio (CN = 1.23, CPA = 1.60). The model included the SOM-derived cluster membership as the sole independent variable, categorizing each CBG based on its profile across multiple social, environmental, and health dimensions. To account for variations in population size across geographic units, the share of children under eighteen was included as an offset variable, enabling the model to estimate rate ratios rather than raw counts. The model was structured as follows: $$\:{Y}_{i}=NB({\mu\:}_{i},\theta\:)$$ where \(\:{Y}_{i}\:\) is the count of CN or CPA, \(\:{\mu\:}_{i}\) is the expected count, and \(\:\theta\:\) is the overdispersion parameter. The log-link function applied to the data is given by: $$\:\text{log}\left(E\left({Y}_{i}\right)\right)={\beta\:}_{0}+{\beta\:}_{1}{SOM}_{i}+\text{l}\text{o}\text{g}\left(pop\right)$$ where log( pop ) is the offset term used to normalize the CN and CPA counts within each CBG, and \(\:{\beta\:}_{1}\) is the effect of SOM cluster membership on the rate of CN and CPA. To quantify differences between clusters, rate ratios (RRs) were computed along with 95% confidence intervals (CIs). A rate ratio greater than one indicates a higher rate of neglect or physical abuse in the comparison cluster relative to the reference cluster. To reduce the risk of Type I error, we applied a correction for multiple comparisons, ensuring that statistically significant results remained robust. The regression model was conducted using the MASS package's glm.nb function in R v 4.4.2. Results Descriptive Results. The correlation analysis reveals significant relationships between tree equity, park priority, green space, and socioeconomic indicators (Fig. 3 ). Tree equity is positively correlated with green space (%) ( r = 0.69, p < 0.001) and tree cover (%) ( r = 0.73, p < 0.001), while negatively associated with the percent of non-White persons ( r = -0.76, p < 0.001) and park need ( r = -0.67, p < 0.001). Park priority is positively correlated with the percentage of poverty ( r = 0.69, p < 0.001), greater physical inactivity (r = 0.66, p < 0.001), and higher mental health concerns (r = 0.53, p < 0.001). Other notable correlations include low physical activity and poverty ( r = 0.55, p < 0.001) and mental health issues and poverty ( r = 0.44, p < 0.001). Green space has a negative correlation with park need ( r = -0.62, p < 0.001). While traditional correlation analysis identifies pairwise relationships, the component planes plot (Fig. 4 ) reveals the co-occurrence of social, physical, and environmental characteristics across geographic areas. As expected, TES, tree cover, and green space exhibit similar distributions, indicating that areas with higher tree coverage and green space also have higher tree equity scores. In contrast, an inverse correlation between park need and park priority reflects the increased park need in areas dominated by historically marginalized groups. The share of the non-White population exhibits a similar spatial clustering pattern to unemployment and poverty, with higher values concentrated in the same regions. Poor mental health and low physical activity also overlap, suggesting that areas with poor mental health tend to have lower physical activity levels. Temperature anomaly aligns with areas of lower tree cover and green space, indicating that regions with less green infrastructure experience more significant temperature fluctuations. The component planes plot also reveals the interrelationship between environmental, physical, and social factors. For example, low TES overlap with areas with high intersection density, high poverty rates, high unemployment, and a greater proportion of non-White residents. Regions with greater intersection density also tend to align with higher poverty and unemployment rates, as well as a greater percentage of non-White residents. In contrast, areas with higher TES, which indicate better access to urban tree coverage, tend to have lower intersection density, poverty, and unemployment rates. These areas also have fewer non-White residents, highlighting disparities in the distribution of tree canopies and access to green infrastructure across different social and economic groups. Table 2 Descriptive characteristics of socio-environmental characteristics within Self-Organizing Map clusters Cluster 1: Urban, High Vulnerability, Low Park, and Tree Equity (N = 84; 3.2%) Cluster 2: Urban, High Vulnerability, High Park Need, Moderate Tree Equity (N = 225; 8.6%) Cluster 3: Suburban, Low Vulnerability, Moderate Park Need, High Tree Equity (N = 544; 20.9%) Cluster 4: Urban, Mixed Vulnerability, High Park Need, Moderate Tree Equity (N = 385; 14.8%) Cluster 5: Suburban, Moderate Vulnerability, Moderate Park Need, High Tree Equity (N = 391; 15.0%) Cluster 6: Suburban, Low Vulnerability, Low Park Need, High Tree Equity (N = 470; 18.1%) Cluster 7: Suburban, Lowest Vulnerability, Lowest Park Need, Highest Tree (N = 504; 19.4%) Tree Equity Score 64.78 (8.09) 68.98 (8.26) 91.18 (8.19) 82.93 (8.07) 74.06 (9.16) 86.30 (9.00) 77.67 (6.68) Park Priority 3.69 (0.53) 3.50 (0.53) 1.89 (0.61) 2.63 (0.57) 3.07 (0.50) 2.10 (0.41) 2.66 (0.50) Green Space 20.37 (7.23) 24.02 (7.18) 40.17 (10.57) 26.44 (9.58) 25.12 (8.41) 35.52 (7.48) 29.35 (5.53) Tree Canopy 19.90 (7.80) 21.88 (6.28) 36.50 (11.50) 28.58 (10.37) 22.92 (8.17) 31.34 (8.69) 28.93 (6.13) Park Need 3.69 (0.53) 3.50 (0.53) 1.89 (0.61) 2.63 (0.57) 3.07 (0.50) 2.10 (0.41) 2.66 (0.50) Park Access 0.23 (0.31) 0.30 (0.27) 0.34 (0.31) 0.27 (0.34) 0.28 (0.27) 0.50 (0.26) 0.53 (0.30) Intersection Density 59.01 (13.36) 65.32 (11.61) 62.51 (24.40) 65.51 (13.69) 65.03 (10.73) 56.46 (12.79) 53.29 (10.36) Pop/Acre 42.52 (21.04) 38.92 (19.45) 20.47 (17.35) 40.83 (23.24) 37.11 (20.70) 16.14 (7.78) 22.11 (11.29) Unemployment 0.09 (0.07) 0.09 (0.06) 0.06 (0.06) 0.09 (0.08) 0.07 (0.05) 0.07 (0.06) 0.06 (0.05) Dependency Ratio 0.58 (0.19) 0.52 (0.19) 0.60 (0.37) 0.28 (0.24) 0.49 (0.25) 0.58 (0.23) 0.53 (0.21) Low Physical Activity 28.09 (7.67) 25.23 (7.49) 11.70 (4.91) 5.24 (6.68) 21.15 (7.90) 15.03 (5.54) 18.85 (8.05) Person of Color 0.93 (0.12) 0.90 (0.14) 0.39 (0.20) 0.50 (0.23) 0.77 (0.21) 0.53 (0.25) 0.74 (0.21) Poor Mental Health 18.00 (4.55) 16.44 (4.28) 10.76 (3.73) 5.30 (6.40) 15.08 (4.37) 12.20 (3.55) 13.46 (5.26) Poverty 0.55 (0.16) 0.47 (0.16) 0.16 (0.12) 0.31 (0.19) 0.40 (0.20) 0.19 (0.12) 0.31 (0.15) Children 0.27 (0.07) 0.22 (0.09) 0.16 (0.09) 0.09 (0.08) 0.19 (0.10) 0.18 (0.07) 0.20 (0.08) The Self-Organizing Map (SOM) analysis identified distinct neighborhood clusters based on environmental, social, and health indicators. Table 2 presents the mean values of each variable within the clusters. Figure 5 maps the SOM clusters across the city. Cluster 1: Urban, High Vulnerability, Low Park, and Tree Equity (N = 84; 3.2%). Cluster 1 represents the highest social, health, and environmental vulnerability levels in densely populated urban areas. These neighborhoods have low green space coverage (20.37%), limited tree canopy (19.90%), and the lowest tree equity score (64.78). Park need is very high (3.69), and only 23% of residents can access a park within a 10-minute walk. The built environment is dense, with high intersection density (59.01) and a population density of 42.52 people per acre. Socioeconomic conditions reflect severe disadvantage, with relatively high unemployment (9%), elevated poverty (55%), and a high dependency ratio (0.58). The population is predominantly people of color (93%), with low physical activity levels (28.09%) and the highest prevalence of poor mental health (18.00%). Children make up 27% of the population. Cluster 2: Urban, High Vulnerability, High Park Need, Moderate Tree Equity (N = 225; 8.6%). Cluster 2 shares many characteristics with Cluster 1, including high social and health vulnerability, but has slightly better environmental conditions. Tree equity (68.98), green space (24.02%), and tree canopy (21.88%) are somewhat improved. Park need remains high (3.50), with 30% of residents living more than a 10-minute walk from a park. The area is highly urbanized, with an intersection density of 65.32 and a population density of 38.92 people per acre. Unemployment (9%), poverty (47%), and the percentage of residents of color (90%) remain high. Physical activity is low (25.23%), and poor mental health is prevalent (16.44%). Children comprise 22% of the population. Cluster 3: Suburban, Low Vulnerability, Moderate Park Need, High Tree Equity (N = 544; 20.9%). Cluster 3 reflects low social vulnerability and moderate health and environmental vulnerability levels in suburban areas. Tree equity is high (91.18%), with green space (40.17%) and tree canopy coverage (36.50%) among the highest. Park need is low (1.89), although 34% of residents lack access to a park within a 10-minute walk. The built environment is moderately dense, with an intersection density of 62.51 and a population density of 20.47 per acre. Unemployment (6%) and poverty (16%) are low, and the population includes fewer people of color (39%). Physical activity levels are higher, and poor mental health (10.76%) is lower than in more vulnerable clusters. Children make up 16% of the population. Cluster 4: Urban, Mixed Vulnerability, High Park Need, Moderate Tree Equity (N = 385; 14.8%). Cluster 4 has mixed levels of vulnerability, combining urban density (40.83 people per acre) with moderate environmental conditions: tree equity (82.93), green space (26.44%), and tree canopy (28.58%). Park need remains high (2.63), and only 27% of residents live within a 10-minute walk of a park. The population is 50% people of color. Despite having a high intersection density (65.51), this cluster has the lowest levels of physical inactivity (5.24%) and the lowest poor mental health prevalence (5.30%). However, social conditions vary: poverty is moderate (31%), unemployment is high (9%), and the dependency ratio is low (0.28). Children make up 9% of the population. Cluster 5: Suburban, Moderate Vulnerability, Moderate Park Need, High Tree Equity (N = 391; 15.0%). Cluster 5 reflects moderate social, health, and environmental vulnerability. Tree equity (74.06), green space (25.12%), and tree canopy (22.92%) are moderate. Park need is elevated (3.07), and only 28% of residents have nearby park access. The area is moderately dense (37.11 people per acre; intersection density 65.03). Poverty (40%) and unemployment (7%) suggest moderate disadvantage. Residents are 77% people of color, and poor mental health (15.08%) and physical inactivity (21.15%) are relatively high. Children make up 19% of the population. Cluster 6: Suburban, Low Park Need, Low Social Vulnerability, Moderate Health Burden (N = 470; 18.1%). Cluster 6 demonstrates low social and environmental vulnerability, with some moderate health concerns. Tree equity (86.30), green space (35.52%), and tree canopy (31.34%) are high. Park need is low (2.10), and 50% of residents have park access, which is one of the highest levels. It is a lower-density area (16.14 people per acre; intersection density 56.46). Socially, this cluster shows low poverty (19%), unemployment (7%), and a moderate racial composition (53% people of color). Physical inactivity (15.03%) and poor mental health (12.20%) suggest relatively lower levels of physical activity and poorer mental health compared to clusters 3 and 4, but higher than clusters 1 and 2. Children make up 18% of the population. Cluster 7: Suburban, Lowest Vulnerability, Lowest Park Need, Highest Tree Equity (N = 504; 19.4%). Cluster 7 represents the least vulnerable cluster overall. It has the highest tree equity (94.98%), the most green space (46.10%), and the highest tree canopy (41.25%). Park need is lowest (1.43), and 53% of residents have park access, the highest across clusters. This cluster is also the least dense, with a population density of 11.71 people per acre and the lowest intersection density (53.29). It has the lowest poverty (12%), lowest unemployment (6%), and lowest percentage of people of color (28%). Physical inactivity (10.70%) and poor mental health (9.97%) are low. Children comprise 20% of the population. Table 3 simplifies the clusters by social, health and environmental vulnerability. Cluster Social Vulnerability Health Vulnerability Environmental Vulnerability Population Density Cluster Summary Cluster 1: Urban, High Vulnerability, Low Park, and Tree Equity (N = 84; 3.2%) ★★★ ★★★ ★★★ ★★★ Highest social disadvantage, poor health outcomes, degraded environmental conditions, and extreme urban density. Cluster 2: Urban, High Vulnerability, High Park Need, Moderate Tree Equity (N = 225; 8.6%) ★★★ ★★★ ★★ ★★★ High social and health vulnerability with slightly better environmental features than Cluster 1, but still dense and disadvantaged. Cluster 3: Suburban, Low Vulnerability, Moderate Park Need, High Tree Equity (N = 544; 20.9%) ★ ★★ ★★ ★★ Low social vulnerability and moderate health burden; strong tree equity but limited park access and moderate density Cluster 4: Urban, Mixed Vulnerability, High Park Need, Moderate Tree Equity (N = 385; 14.8%) ★★ ★ ★★ ★★ Moderate levels of social and health vulnerability, with persistent park need and urban form. Cluster 5: Suburban, Moderate Vulnerability, Moderate Park Need, High Tree Equity (N = 391; 15%) ★★ ★★ ★★ ★★ Moderate social and health burden; good tree equity but barriers to park access remain in a moderately dense setting Cluster 6: Suburban, Low Vulnerability, Low Park Need, High Tree Equity (N = 470; 18.1%) ★ ★★ ★★ ★ Low social vulnerability and low population density; health and environmental conditions are favorable but not optimal. Cluster 7: Suburban, Lowest Vulnerability, Lowest Park Need, Highest Tree (N = 504; 19.4%) ★ ★ ★ ★ Lowest vulnerability across all domains—social, health, environmental, and density—serves as the reference group Notes. Summary of the clusters with three different versions of stars to indicate high (★★★), moderate (★★), or low (★) for each category: social, health, environmental, and density. Social vulnerability : Based on the average of people of color, unemployment, and poverty rates. Health : Based on the average of low physical activity and poor mental health rates. Environmental : Based on the average of tree equity, green space, and tree canopy, park need, proximity to green space, temperature anomalies, and intersection density. Density : Based on the population per acre. The SOM clustering visualization (Fig. 6 ) reveals important patterns in socio-environmental conditions across neighborhoods. The legend shows the multiple variables incorporated into this analysis, ranging from environmental factors (TES, park access, extreme heat) to socioeconomic indicators (unemployment, poverty) and demographic characteristics (dependency ratio, non-white population). The hexagonal grid organizes high-dimensional data into a two-dimensional representation where similar neighborhoods are positioned closer together. The black boundary lines between clusters highlight transition zones where neighborhood characteristics change significantly. These boundaries indicate areas of sharp differentiation in the underlying variables, including park access, extreme heat exposure, population density, socioeconomic factors, and health metrics. Clusters 3, 5, 6, and 7 occupy larger portions of the SOM, indicating these neighborhood types are more prevalent throughout the city. Cluster 2 (yellow) appears centrally located and tightly packed, suggesting a distinct neighborhood typology with consistent characteristics. Clusters 5 (blue) and 6 (purple) are adjacent on the map, indicating a gradual transition between their environmental and social characteristics. Similarly, Cluster 3 (green) and Cluster 7 (pink/magenta) occupy the right portion of the map, showing they share certain similarities while maintaining distinct profiles. Negative Binomial Regression Results for Child Neglect and Physical Abuse The negative binomial regression model indicated statistically significant differences in CN and CPA rates across the SOM-defined neighborhood clusters (CN: Omnibus test X ² = 15.084, df = 5, p = 0.020; CPA: X ² = 26.103, df = 5, p < 0.001). For CN, neighborhoods in Cluster 1 (Urban, High Vulnerability, Low Park, and Tree Equity) had the highest rates of child neglect. Compared to Cluster 1, CN rates were significantly lower in Cluster 4 (Urban, Mixed Vulnerability, High Park Need, Moderate Tree Equity) (IRR = 0.54, p < 0.05) and Cluster 7 (Suburban, Lowest Vulnerability, Lowest Park Need, Highest Tree Equity) (IRR = 0.46, p < 0.05), indicating a strong gradient in risk aligned with social and environmental disadvantage. A similar pattern emerged for CPA: rates were highest in Cluster 1, with CPA incidence 2.19 times higher than in Cluster 7 ( p < 0.001) and 1.97 times higher than in Cluster 5 (Suburban, Moderate Vulnerability, Moderate Park Need, High Tree Equity) ( p < 0.001). In contrast, Cluster 3 (Suburban, Low Vulnerability, Moderate Park Need, High Tree Equity) had significantly lower CPA rates than Cluster 1 (IRR = 0.27, p < 0.05), further supporting the association between socio-environmental advantage and reduced maltreatment risk. To explore how these patterns relate to structural disadvantage, we plotted CN and CPA rates against poverty and racial composition across neighborhoods, grouped by SOM cluster. The scatterplots in Fig. 7 help visualize how rates of CN and CPA vary jointly with poverty levels and the percentage of non-White residents within each cluster. As shown by the figure, higher CN and CPA rates are concentrated across both panels in the upper-right quadrant, corresponding to neighborhoods with relatively higher poverty levels and proportions of non-White residents. This trend is particularly evident in clusters characterized by high social and environmental vulnerability, where maltreatment rates are not only elevated but also more densely concentrated. This visualization complements the results of the negative binomial regression by illustrating that CN and CPA are shaped by the combined burden of racialized economic disadvantage and socioeconomic disadvantage, particularly in urban contexts with limited access to green space and other protective community resources including accessible parks, tree canopy coverage, walkable environments, and protections against extreme heat. Discussion This study applies a socio-ecological lens rooted in environmental justice to highlight the significance of environmental assets in promoting physical transformation and enhancing social well-being. Employing SOMs within the context of environmental justice, this research uncovers patterns among different environmental, social, and health variables drawn from a range of georeferenced datasets. Through the integration of spatial distances and hierarchical clustering, the analysis revealed seven unique, geographically cohesive clusters that showcase diverse neighborhood conditions. These clusters were then used to predict CN and CPA risk. Incorporating spatial proximity ensured that the clusters were both data-driven and spatially meaningful, allowing for a more accurate assessment of the geographic distribution of risk factors. The findings revealed significant correlations between low equity in parks and trees within urban areas and higher rates of CPA and CN. These results are consistent with Green (2022), who underscores the multiscalar nature of child maltreatment risk, where neighborhood-level environmental conditions, such as green space access and urban density, intersect with broader structural and social disadvantage to shape child welfare outcomes (Green, 2022). Together, these findings highlight the need for place-based, equity-focused prevention strategies that address the physical and social dimensions of risk across geodemographic groups. Previous research examining physical and social neighborhood characteristics on CN and CPA rates has yielded mixed results. While prior work found that certain physical conditions (e.g., abandoned dwellings, physical barriers) were associated with reduced maltreatment, other features, such as better-maintained streets, residential decorations, and adequate public transit, were protective for child maltreatment (McDonell & Skosireva, 2009). These divergent findings may reflect limitations of traditional modeling approaches that assume linear and independent relationships between predictors and outcomes. In contrast, the present study used SOMs to capture complex, nonlinear interactions among environmental, social, and health-related indicators of risk. Rather than evaluating variables independently, the SOM approach grouped neighborhoods into clusters based on multidimensional similarity across multiple indicators, producing a typology that more accurately reflects the interaction between socio-environmental vulnerabilities. Our findings align with prior research demonstrating an inverse relationship between green space equity, measured through park accessibility, tree canopy coverage, and equity scores, and socioeconomic vulnerability (Cheruvalath et al., 2022; Liu et al., 2021). Previous studies have similarly found that neighborhoods with higher poverty levels, unemployment, and racialized disadvantage are more likely to suffer from poor environmental quality and reduced access to green infrastructure (Heo et al., 2021). Families living in these under-resourced areas often experience heightened parenting stress due to chronic exposure to both environmental hazards and structural disadvantage (Balseviciene et al., 2014). These stressors—compounded by limited access to green space, urban heat exposure, and high traffic or intersection density—can exacerbate physical and mental health challenges, diminishing caregiver capacity and increasing the risk of child maltreatment (Astell-Burt & Feng, 2019; Balseviciene et al., 2014). Greening interventions in high-risk communities can reduce these burdens by promoting emotional regulation, enhancing physical activity, fostering positive social interactions, and improving parental well-being. However, anti-poverty strategies alone are insufficient to address the broad range of risk factors associated with CPA and CN. Policies must also address the unequal distribution of environmental resources within socioeconomically vulnerable areas. Without explicit attention to disparities in the built environment, structural inequalities will likely persist or worsen, limiting the effectiveness of otherwise well-intentioned public health interventions. An environmental justice approach prioritizes equity in the distribution of environmental resources while addressing the intersectional impacts of social and environmental disadvantage. Effective interventions must be comprehensive, addressing individual, family-level, and structural barriers such as insufficient green infrastructure and uneven urban development that have disproportionately affected communities of color. Targeted investments in historically underserved neighborhoods are essential for tackling structural injustices and promoting the well-being of children and their families. One example of such an initiative is the Green Community Mapping Project in Kalamazoo, which aims to identify high-priority areas for expanding access to green spaces (The Trust for Public, n.d.; Tuddenham, n.d.). Through a partnership among the Children and Nature Network, Trust for Public Land, and the Kellogg Foundation, the project has identified socioeconomically vulnerable neighborhoods where children lack access to green spaces within a ten-minute walk. The project’s findings have been used to guide strategic urban planning, enabling city agencies and community organizations to embed green space development within broader health promotion strategies. Although the project primarily targets physical health outcomes, such as reducing obesity and improving overall well-being, our study highlights its relevance for violence prevention. Specifically, our findings demonstrated that the highest rates of CN and CPA were concentrated in neighborhoods with high social disadvantage and environmental deprivation—areas where access to green space and tree equity were the lowest. Initiatives like the Green Community Mapping Project are consistent with our results in prioritizing place-based interventions in communities where intersecting vulnerabilities are greatest. By expanding green space access in high-risk areas, such efforts have the potential to reduce parenting stress, improve mental health, and strengthen community networks to mitigate child maltreatment risk. Limitations This study contributes new insights to the literature on child welfare by utilizing SOMs as a means to capture better geospatial influences on the intersection of neighborhood and individual factors that increase the risk of child maltreatment. However, several limitations should be considered when interpreting these findings. First, we analyze this data within specific geographic areas, particularly across Los Angeles, which is uniquely situated in terms of its landscape, resources, and political environment. Therefore, the findings from this study may not be applicable for predicting risk factors related to child maltreatment throughout the United States. Similarly, Los Angeles has a distinct climate that affects its ability to generate green space, which may be limited in other regions of California. Another limitation is that the outcome is not defined by residential location but by the event's location; thus, there is an assumption that the event occurred in the residential neighborhood, which we cannot confirm with certainty. Additionally, this study found that the most prominent cluster included neighborhoods with higher tree density and green space, alongside elevated levels of employment, low levels of poverty, high rates of physical activity, and low incidences of poor mental health; however, we cannot establish a causal link between green space and the presence of trees and these individual resilience factors based on this cross-sectional analysis. To better model these complex interactions, future research should continue to disentangle how intersecting neighborhood conditions structure child maltreatment risk, for example, using higher-order random effects models such as Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA) (C. R. Evans et al., 2024). These findings highlight the need to contextualize neighborhood features within broader ecological systems and support integrated, place-based interventions that move beyond surface-level infrastructure improvements or narrowly focused anti-poverty strategies. Conclusion By incorporating green space equity, defined as the equitable distribution of green resources in the most socially vulnerable neighborhoods, this study advances our understanding of how environmental justice can mitigate CN and CPA risk. Targeting built environment interventions in communities with limited access to parks, green spaces, and recreational resources is essential for protecting children and promoting safety. Declarations As this study involved the secondary analysis of publicly available, de-identified data, it was not considered as human subjects research, as defined by federal regulations, and did not require approval from an ethics committee or Institutional Review Board (IRB). Funding: The authors have no financial or proprietary interests in any material discussed in this article. Author Contribution GEBS led the conceptualization, methodology design, and data curation for the study. GEBS also wrote the main text, supervised the project, and contributed to editing and revision. SD contributed to the conceptualization of the study and writing of the manuscript. BR was involved in conceptualization, writing, and editing. CS contributed to writing and editing the manuscript. All authors reviewed and approved the final version of the manuscript. Data Availability All data used in this study are publicly available from open-access sources. The data can be accessed through the following links: Child Physical Abuse and Neglect: City of Los Angeles Open Data Portal, Crime Data from 2020 to Presenthttps://data.lacity.org/Public-Safety/Crime-Data-from-2020-to-Present/2nrs-mtv8. Tree Equity Score is available from American Forests at https://www.americanforests.org/initiatives/tree-equity/. The heat extremity & Park Equity Score can be downloaded from the Trust for Public Land’s ParkServe Database at https://parkserve.tpl.org. The U.S. Environmental Protection Agency, EnviroAtlas contains the data on green space along walkable road networks and can be downloaded at https://www.epa.gov/enviroatlas/enviroatlas-data. Physical inactivity and poor mental health was curated from the CDC PLACES Project (https://www.cdc.gov/places/) and is also available from the American Forests database. The socioeconomic indicators can be downloaded from the U.S. Census Bureau, American Community Surveyhttps://www.census.gov/programs-surveys/acs. All datasets are openly accessible, and no special access or permissions are required. References Andrienko, G., Andrienko, N., Bremm, S., Schreck, T., Von Landesberger, T., Bak, P., & Keim, D. (2010). Space‐in‐Time and Time‐in‐Space Self‐Organizing Maps for Exploring Spatiotemporal Patterns. Computer Graphics Forum , 29 (3), 913–922. https://doi.org/10.1111/j.1467-8659.2009.01664.x Astell-Burt, T., & Feng, X. (2019). Association of Urban Green Space With Mental Health and General Health Among Adults in Australia. JAMA Network Open , 2 (7), e198209. https://doi.org/10.1001/jamanetworkopen.2019.8209 Augustijn, E.-W., & Zurita-Milla, R. (2013). Self-organizing maps as an approach to exploring spatiotemporal diffusion patterns. International Journal of Health Geographics , 12 (1), 60. https://doi.org/10.1186/1476-072X-12-60 Azar, S. T., & Weinzierl, K. M. (2005). Child Maltreatment and Childhood Injury Research: A Cognitive Behavioral Approach. Journal of Pediatric Psychology , 30 (7), 598–614. https://doi.org/10.1093/jpepsy/jsi046 Balseviciene, B., Sinkariova, L., & Andrusaityte, S. (2014). Do green spaces matter? The associations between parenting stress, child mental health problems and green spaces. Procedia-Social and Behavioral Sciences , 140 , 511–516. Barboza, G., Angulski, K., Hines, L., & Brown, P. (2022). Variability in Opioid-Related Drug Overdoses, Social Distancing, and Area-Level Deprivation during the COVID-19 Pandemic: A Bayesian Spatiotemporal Analysis. Journal of Urban Health , 99 (5), 873–886. Barboza, G. E., & Schiamberg, L. (2021). Dual trajectories of parenting self‐efficacy and depressive symptoms in new, postpartum mothers and socioemotional adjustment in early childhood: A growth mixture model. Infant Mental Health Journal , 42 (5), 636–654. https://doi.org/10.1002/imhj.21937 Barboza-Salerno, G. E. (2020). Cognitive readiness to parent, stability and change in postpartum parenting stress and social-emotional problems in early childhood: A second order growth curve model. Children and Youth Services Review , 113 , 104958. https://doi.org/10.1016/j.childyouth.2020.104958 Barboza-Salerno, G. E. (2023). The neighborhood deprivation gradient and child physical abuse and neglect: A Bayesian spatial model. Child Abuse & Neglect , 146 , 106501. Barboza-Salerno, G. E., & Meshelemiah, J. C. (2023). Gun Violence on Walkable Routes to and from School: Recommendations for Policy and Practice. Journal of Urban Health , 1–16. Basara, H. G., & Yuan, M. (2008). Community health assessment using self-organizing maps and geographic information systems. International Journal of Health Geographics , 7 (1), 67. https://doi.org/10.1186/1476-072X-7-67 Bywaters, P., Brady, G., Sparks, T., & Bos, E. (2016). Child welfare inequalities: New evidence, further questions. Child & Family Social Work , 21 (3), 369–380. https://doi.org/10.1111/cfs.12154 Cause, A. G. (2020). The Economically Disadvantaged Speak: Exploring the Intersection of Poverty, Race, Child Neglect and Racial Disproportionality in the Child Welfare System [PhD Thesis, Portland State University]. https://search.proquest.com/openview/9105306031fcc5ec3764f7575d7fbf4b/1?pq-origsite=gscholar&cbl=18750&diss=y&casa_token=8hpVkVmiqMIAAAAA:Nvwc3mdUlWgnBmtRZWMOl4foacEuhB2c __1YMnt9BdBasQvqMTrB1Sox7rgUXvD6RqfDsfrLGoM Chen, W., Wu, A. N., & Biljecki, F. (2021). Classification of urban morphology with deep learning: Application on urban vitality. Computers, Environment and Urban Systems , 90 , 101706. https://doi.org/10.1016/j.compenvurbsys.2021.101706 Cheruvalath, H., Homa, J., Singh, M., Vilar, P., Kassam, A., & Rovin, R. A. (2022). Associations Between Residential Greenspace, Socioeconomic Status, and Stroke: A Matched Case-Control Study. Journal of Patient-Centered Research and Reviews , 9 (2), 89–97. https://doi.org/10.17294/2330-0698.1886 Chung, G., Lanier, P., & Wong, P. Y. J. (2022). Mediating Effects of Parental Stress on Harsh Parenting and Parent-Child Relationship during Coronavirus (COVID-19) Pandemic in Singapore. Journal of Family Violence , 37 (5), 801–812. https://doi.org/10.1007/s10896-020-00200-1 Elise Barboza-Salerno, G. (2024). Material Hardship, Labor Market Characteristics and Substantiated Child Maltreatment: A Bayesian Spatiotemporal Analysis. Children and Youth Services Review , 157 , 107371. https://doi.org/10.1016/j.childyouth.2023.107371 Ellis, W. E., Dumas, T. M., & Forbes, L. M. (2020). Physically isolated but socially connected: Psychological adjustment and stress among adolescents during the initial COVID-19 crisis. Canadian Journal of Behavioural Science/Revue Canadienne Des Sciences Du Comportement , 52 (3), 177. Evans, C. R., Leckie, G., Subramanian, S. V., Bell, A., & Merlo, J. (2024). A tutorial for conducting intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA). SSM - Population Health , 26 , 101664. https://doi.org/10.1016/j.ssmph.2024.101664 Evans, M. F., Gazze, L., & Schaller, J. (2025). Temperature and maltreatment of young children. Review of Economics and Statistics , 1–37. Fan, Y., Das, K. V., & Chen, Q. (2011). Neighborhood green, social support, physical activity, and stress: Assessing the cumulative impact. Health & Place , 17 (6), 1202–1211. Finno-Velasquez, M., He, A. S., Perrigo, J. L., & Hurlburt, M. S. (2017). Community Informant Explanations for Unusual Neighborhood Rates of Child Maltreatment Reports. Child and Adolescent Social Work Journal , 34 (3), 191–204. https://doi.org/10.1007/s10560-016-0463-3 Font, S. A., & Maguire-Jack, K. (2015). Decision-making in child protective services: Influences at multiple levels of the social ecology. Child Abuse & Neglect , 47 , 70–82. https://doi.org/10.1016/j.chiabu.2015.02.005 Foundation, T. A. E. C. (2021, October 12). Child Maltreatment Trends . The Annie E. Casey Foundation. https://www.aecf.org/blog/child-maltreatment-trends Freisthler, B. (2004). A spatial analysis of social disorganization, alcohol access, and rates of child maltreatment in neighborhoods. Children and Youth Services Review , 26 (9), 803–819. Freisthler, B., Gruenewald, P. J., Treno, A. J., & Lee, J. (2003). Evaluating alcohol access and the alcohol environment in neighborhood areas. Alcoholism: Clinical and Experimental Research , 27 (3), 477–484. Frumkin, H., Bratman, G. N., Breslow, S. J., Cochran, B., Kahn Jr, P. H., Lawler, J. J., Levin, P. S., Tandon, P. S., Varanasi, U., Wolf, K. L., & Wood, S. A. (2017). Nature Contact and Human Health: A Research Agenda. Environmental Health Perspectives , 125 (7), 075001. https://doi.org/10.1289/EHP1663 Green, J. W. (2022). The Built Environment and Predicting Child Maltreatment: An Application of Random Forests to Risk Terrain Modeling. The Professional Geographer , 74 (1), 67–78. https://doi.org/10.1080/00330124.2021.1970591 Guterman, N. B., Lee, S. J., Taylor, C. A., & Rathouz, P. J. (2009). Parental perceptions of neighborhood processes, stress, personal control, and risk for physical child abuse and neglect. Child Abuse & Neglect , 33 (12), 897–906. Haas, B. M., Berg, K. A., Schmidt-Sane, M. M., Korbin, J. E., & Spilsbury, J. C. (2018). How might neighborhood built environment influence child maltreatment? Caregiver perceptions. Social Science & Medicine , 214 , 171–178. https://doi.org/10.1016/j.socscimed.2018.08.033 Heo, S., Nori-Sarma, A., Kim, S., Lee, J.-T., & Bell, M. L. (2021). Do persons with low socioeconomic status have less access to greenspace? Application of accessibility index to urban parks in Seoul, South Korea. Environmental Research Letters , 16 (8), 084027. Kendrick, E. (2015). A Neighborhood Study: Recreational Parks and Parent Stress [PhD Thesis, The Ohio State University]. https://kb.osu.edu/items/ffeb5391-0a3d-5081-b00f-2cf781f250b4 Kohonen, T. (1990). The self-organizing map. Proceedings of the IEEE , 78 (9), 1464–1480. Kuo, F. E., & Sullivan, W. C. (2001a). Aggression and Violence in the Inner City: Effects of Environment via Mental Fatigue. Environment and Behavior , 33 (4), 543–571. https://doi.org/10.1177/00139160121973124 Kuo, F. E., & Sullivan, W. C. (2001b). Environment and Crime in the Inner City: Does Vegetation Reduce Crime? Environment and Behavior , 33 (3), 343–367. https://doi.org/10.1177/0013916501333002 Le, K. (2025). The Impacts of Extreme Heat Days on the Prevalence of Domestic Abuse. Sage Open , 15 (1), 21582440251317797. https://doi.org/10.1177/21582440251317797 Littleton, T., Freisthler, B., Boyd, R., Smith, A. M., & Barboza-Salerno, G. (2024). Historical redlining, neighborhood disadvantage, and reports of child maltreatment in a large urban county. Child Abuse & Neglect , 156 , 107011. https://doi.org/10.1016/j.chiabu.2024.107011 Liu, D., Kwan, M.-P., & Kan, Z. (2021). Analysis of urban green space accessibility and distribution inequity in the City of Chicago. Urban Forestry & Urban Greening , 59 , 127029. Maguire-Jack, K., & Font, S. A. (2017). Community and individual risk factors for physical child abuse and child neglect: Variations by poverty status. Child Maltreatment , 22 (3), 215–226. Maguire-Jack, K., & Showalter, K. (2016). The protective effect of neighborhood social cohesion in child abuse and neglect. Child Abuse & Neglect , 52 , 29–37. Maguire-Jack, K., Yoon, S., & Hong, S. (2022). Social cohesion and informal social control as mediators between neighborhood poverty and child maltreatment. Child Maltreatment , 27 (3), 334–343. McDonell, J., & Skosireva, A. (2009). Neighborhood Characteristics, Child Maltreatment, and Child Injuries. Child Indicators Research , 2 (2), 133–153. https://doi.org/10.1007/s12187-009-9038-6 Morris, A. S., Robinson, L. R., Hays‐Grudo, J., Claussen, A. H., Hartwig, S. A., & Treat, A. E. (2017). Targeting Parenting in Early Childhood: A Public Health Approach to Improve Outcomes for Children Living in Poverty. Child Development , 88 (2), 388–397. https://doi.org/10.1111/cdev.12743 Morton, C. M., Simmel, C., & Peterson, N. A. (2014). Neighborhood alcohol outlet density and rates of child abuse and neglect: Moderating effects of access to substance abuse services. Child Abuse & Neglect , 38 (5), 952–961. Moudon, A. V., Lee, C., Cheadle, A. D., Garvin, C., Johnson, D., Schmid, T. L., Weathers, R. D., & Lin, L. (2006). Operational definitions of walkable neighborhood: Theoretical and empirical insights. Journal of Physical Activity and Health , 3 (s1), S99–S117. New Child Maltreatment Report Finds Child Abuse and Neglect Decreased to a Five-Year Low . (n.d.). Retrieved December 15, 2024, from https://www.acf.hhs.gov/media/press/2023/new-child-maltreatment-report-finds-child-abuse-and-neglect-decreased-five-year Reports of Child Abuse and Neglect, by Race/Ethnicity . (n.d.). Kidsdata.Org. Retrieved December 15, 2024, from https://www.kidsdata.org/topic/3/reported-abuse-race/table#fmt=1217&loc=2,127,347,1763,331,348,336,171,321,345,357,332,324,369,358,362,360,33 7,327,364,356,217,353,328,354,323,352,320,339,334,365,343,330,367,344,355,366,368,265,349 ,361,4,273,59,370,326,333,322,341,338,350,342,329,325,359,351,363,340,335 &tf=110&ch=7,11,8,10,9 Riley, T., Schleimer, J. P., & Jahn, J. L. (2024). Organized abandonment under racial capitalism: Measuring accountable actors of structural racism for public health research and action. Social Science & Medicine , 343 , 116576. https://doi.org/10.1016/j.socscimed.2024.116576 Roof, K., & Oleru, N. (2008). Public Health: Seattle and King County’s Push for the Built Environment. Journal of Environmental Health , 71 (1), 24–27. The Trust for Public. (n.d.). Kalamazoo, Michigan: Connecting Children with Nature . Retrieved March 17, 2025, from https://www.tpl.org/wp-content/uploads/2013/10/convis-kalamazoo.pdf Tuddenham, K. A. (n.d.). Linking Science—Measuring Health Outcomes . Retrieved March 17, 2025, from https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=00c244ce900cb9ed7742efe7f11a959b49c77823 U.S. Department of Health & Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children’s Bureau. (2024). Child Maltreatment . https://www.acf.hhs.gov/sites/default/files/documents/cb/cm2022.pdf Additional Declarations No competing interests reported. Supplementary Files SupplementaryFigure1SOM.docx Cite Share Download PDF Status: Published Journal Publication published 09 Apr, 2026 Read the published version in International Journal on Child Maltreatment: Research, Policy and Practice → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6521185","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":456685228,"identity":"d4f88d4e-404b-46cd-9181-aeb9dfea76fc","order_by":0,"name":"Gia Barboza-Salerno","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABAklEQVRIiWNgGAWjYBACAwYGNiiTuQFMsbFDaMYGwlqgavh4DpCqRU4iAb8Wc/azzx58+GPHIN9+sPFx5Q4bBjbJx4c/8zDYyG44gF2LZU+6ueEMnmQGgzOJzYZnz6QxsEmnpUnzMKQZ49JicCCNTZpHgrl+A0Nim2Rj22GglhwzZh6Gw4k4tZx/BtRiUM8g3/+w/SdYi+QZY6DD/uPWcgNkS8JhBoYbiW2MYC0SPAZAhx3AqcVyxjN2wxkHjgP1PmwGOiyNh40nLU1yjkGy8UwcWsz509iAIVYNdFjywY+NbTZy8u2HD394U2En24dDCwbggTqYSOWjYBSMglEwCrACANlRVk4LsZ8WAAAAAElFTkSuQmCC","orcid":"","institution":"The Ohio State University","correspondingAuthor":true,"prefix":"","firstName":"Gia","middleName":"","lastName":"Barboza-Salerno","suffix":""},{"id":456685229,"identity":"d0b2481a-d483-4384-b9fe-4c101532ef59","order_by":1,"name":"Sharefa Duhaney","email":"","orcid":"","institution":"The Ohio State University","correspondingAuthor":false,"prefix":"","firstName":"Sharefa","middleName":"","lastName":"Duhaney","suffix":""},{"id":456685230,"identity":"abf7a70c-2e22-4c6a-8455-48e79c3b0d40","order_by":2,"name":"Balaji Ramesh","email":"","orcid":"","institution":"The Ohio State University","correspondingAuthor":false,"prefix":"","firstName":"Balaji","middleName":"","lastName":"Ramesh","suffix":""},{"id":456685231,"identity":"398afe5c-378e-41ea-95d7-0fe92f1139f3","order_by":3,"name":"Charis Stanek","email":"","orcid":"","institution":"The Ohio State University","correspondingAuthor":false,"prefix":"","firstName":"Charis","middleName":"","lastName":"Stanek","suffix":""}],"badges":[],"createdAt":"2025-04-24 13:23:23","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6521185/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6521185/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s42448-026-00256-4","type":"published","date":"2026-04-09T15:58:26+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":83016683,"identity":"270b7c60-e2b5-4003-9de0-cd6343311897","added_by":"auto","created_at":"2025-05-19 06:32:54","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":241209,"visible":true,"origin":"","legend":"\u003cp\u003eOur conceptual model of the relationship between social vulnerability and environmental burden.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-6521185/v1/eebb6add3276ce4dea3afa13.png"},{"id":83017563,"identity":"50f8b865-c93f-4a7f-9842-ea1d27c27533","added_by":"auto","created_at":"2025-05-19 06:40:54","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":597128,"visible":true,"origin":"","legend":"\u003cp\u003eTree equity score in census block groups (CBGs) (A) Tree cover along walkable street networks raster image (B) in Los Angeles overlaid onto the boundary of Los Angeles County (grey).\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-6521185/v1/14635c69fe2e19cb269cb5f5.png"},{"id":83016677,"identity":"c308bbe6-0b43-4754-af44-ff4cfc4b896a","added_by":"auto","created_at":"2025-05-19 06:32:54","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":237669,"visible":true,"origin":"","legend":"\u003cp\u003eAssociations between key variables used in the Self-Organizing Map analysis.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-6521185/v1/c155688a31798b8d658178f0.png"},{"id":83017793,"identity":"dd8e500f-1334-407b-8142-efaf6307fb16","added_by":"auto","created_at":"2025-05-19 06:48:54","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":209663,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eComponent planes plot.\u003c/em\u003e The component planes plot helps visualize the SOM input vectors by showing how different neighborhood characteristics cluster within neighborhoods. Each hexagonal cell represents a neuron in the SOM, which groups areas with similar characteristics based on the input data. The color gradient (from yellow to red) reflects the relative values of each variable, with darker reds indicating higher values of that variable and lighter yellows indicating lower values.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-6521185/v1/536d03a06da7e47698bfc91a.png"},{"id":83016696,"identity":"1b601044-f10f-4240-937a-a02e06fe3bfb","added_by":"auto","created_at":"2025-05-19 06:32:54","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":419989,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eSpatial distribution of child neglect and child physical abuse across SOM clusters.\u003c/em\u003e This map displays the spatial distribution of child neglect (circles) and child physical abuse (triangles) incidents overlaid on self-organizing map (SOM) clusters derived from social, environmental, and health-related indicators in Los Angeles. Each color represents a distinct SOM cluster defined by the level of social vulnerability and environmental factors. CN and CPA are most concentrated in clusters characterized by high social vulnerability and low green space equity, particularly in dense urban areas. In contrast, areas with lower vulnerability and more equitable environmental conditions show fewer incidents.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-6521185/v1/c9602781564bfbee8230ba65.png"},{"id":83016688,"identity":"3d659cb4-657f-4e87-857b-60b53d3ed479","added_by":"auto","created_at":"2025-05-19 06:32:54","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":296875,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eSelf-Organizing Map Clusters. \u003c/em\u003eThis hexagonal map represents the topology of neighborhood clusters using multidimensional data from environmental, social, and health-related domains. Each hexagon reflects a neighborhood unit, and colors indicate cluster membership, with numbers (1–7) representing the final hierarchical cluster groupings. Radial bar plots within each hexagon show the relative values of 16 color-coded input variables as shown in the legend.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-6521185/v1/3e24f132a34abfb095182ce4.png"},{"id":83017794,"identity":"cd125363-fcd1-4f40-9c52-c4fd3aa8a1d9","added_by":"auto","created_at":"2025-05-19 06:48:54","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":329454,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cem\u003eRelationship Between Poverty and Racial Composition Within SOM Clusters and Child Neglect and Physical Abuse Rates. Panel \u003c/em\u003e(A) shows the distribution of CN rates, showing the relationship between poverty rates and the percentage of non-White residents across neighborhoods, grouped by self-organizing map (SOM) clusters. Panel (B) depicts child physical abuse (CPA) rates. Each point represents a census tract, colored by SOM cluster and scaled by the respective CN or CPA rate.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-6521185/v1/0ad365c372cb264028e03714.png"},{"id":106809395,"identity":"41c3c72f-334b-4756-9909-79630b6a5a80","added_by":"auto","created_at":"2026-04-13 16:10:18","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3400256,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6521185/v1/bb4270c9-1aa1-45a9-bc72-59a74d893511.pdf"},{"id":83017562,"identity":"7d2c0e4c-1fb0-4927-9d3b-593d4b20c330","added_by":"auto","created_at":"2025-05-19 06:40:54","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":40336,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure1SOM.docx","url":"https://assets-eu.researchsquare.com/files/rs-6521185/v1/e11a0770c4cfa2b1d1e040e6.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Beyond Social Disadvantage: Advancing an Environmental Justice Framework to Address Child Maltreatment Risk","fulltext":[{"header":"Introduction","content":"\u003cp\u003eChild abuse and neglect are significant public health concerns with lasting consequences for both children and communities. In 2021, U.S. Child Protective Services (CPS) investigated over 3\u0026nbsp;million cases of suspected maltreatment, with an estimated 600,000 children confirmed as victims of abuse or neglect (\u003cem\u003eNew Child Maltreatment Report Finds Child Abuse and Neglect Decreased to a Five-Year Low\u003c/em\u003e, n.d.; U.S. Department of Health \u0026amp; Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children\u0026rsquo;s Bureau, 2024). Neglect accounted for the majority of these cases, with 70% of substantiated victims under the age of ten (Foundation, 2021). In Los Angeles County, the focus of this study, African American/Black children experience child abuse and neglect reporting rates that are 3.4 times higher than their White counterparts (\u003cem\u003eReports of Child Abuse and Neglect, by Race/Ethnicity\u003c/em\u003e, n.d.). While systemic inequities, including structural racism, contribute to these disparities, few studies explicitly examine how neighborhood socio-environmental conditions shape child maltreatment risk (Littleton et al., 2024). Even fewer apply frameworks that analyze how structural racism operates through socio-environmental characteristics (Foundation, 2021; Riley et al., 2024).\u003c/p\u003e \u003cp\u003eThe social determinants of health within neighborhoods significantly influence child welfare outcomes through interconnected economic, social, and environmental pathways. Previous research has established a correlation between child abuse and neglect and various social and economic risk factors, including elevated poverty and unemployment rates (Barboza-Salerno, 2024), housing instability (Barboza-Salerno, 2020b; Littleton et al., 2024), restricted access to essential services (Maguire-Jack \u0026amp; Font, 2017), and a high concentration of alcohol and drug outlets (Freisthler, 2004; Freisthler et al., 2022). Although social and economic conditions are pivotal, the built environment can either exacerbate or alleviate these risks, depending on the broader social context. Nevertheless, existing research has predominantly overlooked the interplay between the built environment, such as access to green spaces, walkability, and recreational opportunities, and factors associated with Child Protective Services (CPA) and Child Neglect (CN) risk. A notable exception is a recent study that identified an increase in the risk of child welfare involvement related to tree canopy cover and green space, even when controlling for neighborhood social conditions, including area-level deprivation (He et al., 2024). However, the built environment encompasses a range of distinct features\u0026mdash;including extreme heat (Evans et al., 2025; Le, 2025), morphology (e.g., street layout) (Barboza-Salerno \u0026amp; Meshelemiah, 2023; Chen et al., 2021), and equitable access to trees and parks\u0026mdash;that collectively influence behavioral patterns, movement, and social interactions pertinent to CPA and CN risk. Noah (2015, p. 457) characterizes the failure to integrate environmental and social exposures as \u0026ldquo;spatial polygamy,\u0026rdquo; wherein isolated analyses disregard the cumulative effects of overlapping risk factors.\u003c/p\u003e \u003cp\u003eThe purpose of this study is to examine whether neighborhoods with similar social, economic, and physical environmental characteristics exhibit comparable patterns of child physical abuse and neglect risk in Los Angeles, California. To achieve this, we used multiple indicators of the social, economic, and physical environment to classify areas into distinct typologies using Self-Organizing Maps (SOMs). We then mapped the typologies to geographic space to compare small-area rates of CPA and CN within each cluster. Unlike traditional distance-based clustering models, which rely on predefined assumptions about variable relationships, SOMs employ an unsupervised machine learning algorithm to detect nonlinear patterns and emergent zonal structures. This study develops a new framework for understanding the geographic and contextual factors influencing child victimization risk by incorporating less-explored socio-environmental elements, such as equitable park access, urban heat islands, and intersection density, alongside traditional socioeconomic measures.\u003c/p\u003e\n\u003ch3\u003eAssociation Between Neighborhood Social, Physical, and Environmental Characteristics and Child Abuse and Neglect\u003c/h3\u003e\n\u003cp\u003eA substantial body of research has documented the individual (Azar \u0026amp; Weinzierl, 2005; Cause, 2020; Maguire-Jack \u0026amp; Font, 2017) and contextual factors (Barboza-Salerno, 2023; Elise Barboza-Salerno, 2024; Font \u0026amp; Maguire-Jack, 2015; Freisthler, 2004; Freisthler et al., 2003; Littleton et al., 2024; Maguire-Jack et al., 2022) associated with CPA and CN. At the individual level, the risk is closely linked to family characteristics, including a child\u0026rsquo;s age and gender, the caregiver's mental health, material deprivation, substance misuse, and socioeconomic status. Parenting stress, in particular, has been strongly associated with child maltreatment, often mediated by child behavioral problems and poor socioemotional adjustment (Barboza \u0026amp; Schiamberg, 2021; Barboza-Salerno, 2020a). At the contextual level, similar social factors, such as poverty, material deprivation, and residential instability, also play a critical role in shaping maltreatment risk. Moreover, evidence suggests a graded, dose-response relationship between social vulnerability and child maltreatment, with greater exposure to adverse social conditions correlating with increased risk for violence or abuse (Barboza et al., 2022; Barboza-Salerno, 2023; Bywaters et al., 2016).\u003c/p\u003e \u003cp\u003eResearch suggests that the built environment, defined as \u0026ldquo;the human-made space where people live, work, and recreate on a day-to-day basis (Roof \u0026amp; Oleru, 2008),\u0026rdquo; moderates the socioeconomic structure of neighborhoods with important consequences for child well-being (Morton et al., 2014). The built environment encompasses buildings (residential, commercial, and industrial), infrastructure (roads, bridges, water systems, and energy grids), public spaces (parks, plazas, and playgrounds), and landscapes shaped by urban and rural development. A qualitative study examining caregivers\u0026rsquo; views on the built environment's influence on child maltreatment revealed that abandoned buildings, insufficient green space, and poorly maintained infrastructure correlate with a higher risk of child maltreatment (Haas et al., 2018). Abandoned buildings and deteriorating infrastructure can contribute to neighborhood disorder, increasing parental stress by fostering unsafe conditions, limiting access to community resources, and reducing opportunities for social support (Guterman et al., 2009; Maguire-Jack \u0026amp; Showalter, 2016).\u003c/p\u003e \u003cp\u003eIn neighborhoods with limited green spaces, children face an increased risk of poor health outcomes, such as obesity and respiratory issues, due to the role of reduced vegetation in creating urban heat islands (Moudon et al., 2006). Research has shown that proximity to parks is associated with reduced child welfare involvement, lower parenting stress, increased social interactions, and greater physical activity (Frumkin et al., 2017; Kuo \u0026amp; Sullivan, 2001a, 2001b). Densely populated areas with high rental occupancy and shared living spaces can exacerbate maltreatment risks due to overcrowding, unsafe conditions, and limited supervision. Studies have highlighted how urban sprawl\u0026mdash;characterized by imbalanced housing and employment opportunities\u0026mdash;may heighten parenting stress and negatively impact child outcomes. Access to parks and open spaces fosters social interaction and reduces stress (Fan et al., 2011; Kendrick, 2015; Morris et al., 2017), while overcrowded and unsafe living conditions in high-density areas contribute to increased maltreatment risk (Finno-Velasquez et al., 2017). Furthermore, physical isolation\u0026mdash;often resulting from neglected community spaces or barriers to social engagement\u0026mdash;exacerbates vulnerabilities by limiting access to essential resources and social support networks (Chung et al., 2022; Ellis et al., 2020).\u003c/p\u003e \u003cp\u003eThe benefits of living near green spaces are not uniform but vary across different socioeconomic contexts. Therefore, the effect modification of green space on the association between parenting stress and child well-being depends on the broader social conditions that shape parent-child interactions. For example, a study of 1,468 mothers in Lithuania found that greater distances from parks were associated with worse mental health outcomes but only among parents with lower educational attainment. While green spaces offer potential protective effects, other aspects of the built environment may also exacerbate child physical abuse and neglect risk. Morton et al. (2014) found that areas with easier access to substance abuse services had lower rates of neglect, even after controlling for neighborhood demographics and socioeconomic structure. Further, the presence of substance abuse service facilities moderated the impact of alcohol outlet density on child maltreatment rates. Despite these associations, research on child welfare determinants has been fragmented, often overlooking the complex, non-linear interactions between social, economic, and environmental factors that may play a causal role in child victimization (Balseviciene et al., 2014), with only a handful of studies focusing on the broader environmental characteristics contributing to increased risk (Haas et al., 2018).\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eA theoretical and conceptual model\u003c/h2\u003e \u003cp\u003eSocial disorganization theory has traditionally guided sociological and social work perspectives on child maltreatment risk (Green, 2022). However, recent research indicates that the physical characteristics of the built environment also play a significant role, expanding our understanding beyond just socioeconomic and social psychological factors. The focus on socioeconomic factors related to child well-being, particularly poverty, has created a lack of conceptual clarity about how the built environment may contribute to both child poverty and neighborhood disadvantage. According to SDT, neighborhood social disorder disrupts social cohesion and weakens informal social control, reducing collective efficacy and ultimately contributing to higher levels of crime, violence, and abuse. Although SDT\u0026rsquo;s fundamental assertion that neighborhood environments influence child outcomes is well-supported, emerging research indicates that these effects are primarily driven by systemic forces that shape physical and social well-being (Littleton et al., 2024).\u003c/p\u003e \u003cp\u003eWhile SDT focuses on the consequences of neighborhood disorder, it largely overlooks these systemic factors. Decades of redlining and discriminatory urban planning practices have resulted in neighborhoods where social vulnerability overlaps with environmental risk, compounding stressors that increase the likelihood of child maltreatment (Littleton et al., 2024). Community disinvestment erodes essential social support and resources, particularly in impoverished urban areas. Recognizing these structural forces shifts the focus from neighborhood collective efficacy to systemic harm created and sustained by de jure segregation. On this basis, Ruth Gilmore\u0026rsquo;s organized abandonment theory offers a more compelling lens for understanding how neighborhoods contribute to CPA and CN. According to organized abandonment theory, broader patterns of intentional systemic disinvestment create conditions where children face heightened vulnerability and exposure to violence. Unlike SDT, which places the responsibility for resolving neighborhood social and physical \u0026lsquo;disorganization\u0026rsquo; (renaming the word to organization does not change the theory\u0026rsquo;s assumptions) on residents, organized abandonment theory highlights systemic inequities, such as disparities in green space access, poor housing conditions, and environmental burdens, that require alternative law and policy responses.\u003c/p\u003e \u003cp\u003eUsing organized abandonment as a lens to view the mechanisms linking social and physical conditions to CPA and CN, we introduce the conceptual model shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. In our model, organized abandonment primarily operates through systematic disinvestment, creating neighborhoods vulnerable to social, environmental, and health burdens. In turn, socially vulnerable areas have unequal access to high-quality public spaces due to limited walkability and urban decay. Communities facing socioeconomic disadvantages often experience higher environmental burdens, such as exposure to pollution and a lack of green spaces, which compound risks to children's well-being. Neighborhood vulnerability and green space morphology structure interactions between parents and children in specific contexts.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eOur conceptual model is theoretically grounded in Bronfenbrenner\u0026rsquo;s ecological systems theory and Garbarino\u0026rsquo;s human ecology model, which frames child maltreatment as the result of interacting influences across multiple levels, from individual interactions (microsystem) to broader societal structures (macrosystem). Within this framework, natural environment features (e.g., access to green spaces) are critical components that shape child outcomes through public spaces where children and their parents interact. Equitable access to green spaces, walkable streets, and essential amenities is crucial for positive social interactions. Several potential mechanisms explain how these characteristics may heighten CAN risk. First, green spaces, walkability, and community resources reduce parental stress and foster social connections, thereby minimizing the risks of child maltreatment (Balseviciene et al., 2014). Neighborhoods with abundant green spaces may provide safe environments for children to play, socialize, and engage in physical activity, promoting healthy development and reducing CAN risk (Fan et al., 2011). In contrast, neighborhoods with limited or unsafe parks often confine children indoors for extended periods, which increases parental stress, supervision challenges, and risks of both CPA and CN. Additionally, deteriorating or unsafe public spaces expose children to physical hazards, further heightening maltreatment risks (Guterman et al., 2009). The conceptual model informs our methodological approach by identifying key variables (e.g., park equity, intersection density, physical inactivity, and poverty) that serve as input features for training our machine learning model.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eCurrent Study\u003c/h3\u003e\n\u003cp\u003eThis study employs SOMs, an unsupervised machine-learning technique that identifies nonlinear relationships between socioenvironmental and geographic factors to understand how underlying exposure patterns interact in complex, dynamic systems (Basara \u0026amp; Yuan, 2008). By clustering areas with similar characteristics, SOMs reveal hidden patterns in high-dimensional data, providing a more nuanced understanding of how these areas collectively influence child maltreatment risk. Using SOMs, we construct a typology that links spatial patterns of child maltreatment to specific environmental features. By examining the relative risks of CPA and CN across these clusters, we seek to answer two research questions: (a) What neighborhood typologies arise from grouping the built, natural, and social environments in Los Angeles? Furthermore, (b) Which neighborhood typologies have higher rates of CPA and CN? This approach provides insights into socio-environmental factors that shape CPA and CN risk, highlighting the need for prevention strategies tailored to neighborhood-specific needs.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eStudy Population and Design\u003c/h2\u003e \u003cp\u003eThis study uses incident data from the Los Angeles Police Department on CN and CPA occurring between January 1, 2020, and December 31, 2023. Under California law, child neglect (Penal Code \u0026sect;\u0026nbsp;270 PC) is when a parent or legal guardian willfully (and without lawful excuse) fails to provide necessities such as clothing, food, medicine, and shelter. Child abuse (Penal Code \u0026sect;\u0026nbsp;273d PC) is defined as the willful infliction of cruel or inhuman corporal punishment or causing a traumatic injury (e.g., slapping, punching, or hitting a child). To calculate CPA and CN rates, we obtained 5-year population estimates for individuals aged 0 to 17 within each census block group from the American Community Survey (ACS) for 2018\u0026ndash;2023. We merged environmental data on tree and park equity and access from American Forests via the Tree Equity Score, the California Office of Environmental Health Hazard Assessment, and the Trust for Public Land\u0026rsquo;s ParkServe platform. Additionally, physical and mental health indicators were drawn from the Centers for Disease Control and Prevention\u0026rsquo;s (CDC) PLACES dataset, derived from the Behavioral Risk Factor Surveillance System (BRFSS), to capture neighborhood-level health factors that may be linked to CPA and CN outcomes. Dependency information, including the ratio of adults over 65 to children under 18, linguistic isolation, poverty, race, and unemployment, was provided by the ACS 5-year estimates (see Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eIndicators\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDataset\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMetric Type\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSource\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTimeframe\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChild Physical Abuse and Neglect\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHealth \u0026ndash; primary outcome variable\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCity of Los Angeles Open Data Portal \u0026ndash; Crime Incidents from 2000 to present (Crime Data from 2020 to Present | Los Angeles - Open Data Portal)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTree Equity Score (TES)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEnvironmental Burden\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eAmerican Forests (Tree Equity - American Forests)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHeat Extremity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEnvironmental Burden\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eTrust for Public Land\u0026rsquo;s ParkServe Database (ParkServe - Trust for Public Land)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePark Equity Score (PES)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEnvironmental Burden\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGreen Space along road network (percent green space)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEnvironmental Burden\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eEnviroAtlas (EnviroAtlas Data | US EPA)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e2018\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTree Cover along road network (percent cover)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eEnvironmental Burden\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePhysical Inactivity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHealth Burden\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eU.S. Centers for Disease Control and Prevention (CDC) 500 Cities Project, renamed PLACES as of late 2020 (Centers for Disease Control and Prevention et al., 2023)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePoor Mental Health\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHealth Burden\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2023\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge Dependency\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSocial Burden\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\" morerows=\"5\" rowspan=\"6\"\u003e \u003cp\u003eU.S. American Community Survey (ACS) (U.S. Census Bureau, 2023)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2019\u0026ndash;2023\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnemployment (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSocial Burden\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2019\u0026ndash;2023\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePoverty (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSocial Burden\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2019\u0026ndash;2023\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLinguistic Isolation (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSocial Burden\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2019\u0026ndash;2023\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNon-White Residents (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSocial Burden\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2019\u0026ndash;2023\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePercent of Children (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSocial Burden\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2019\u0026ndash;2023\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eEnvironmental burden.\u003c/b\u003e \u003cem\u003eTree- and Park Equity\u003c/em\u003e. The tree equity score (TES; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA) developed by American Forests is used here to evaluate the impact of the distribution of urban tree canopy on CPA and CN risk. The TES is constructed by considering tree canopy goals for each area and identifying priority areas associated with the greatest need for tree coverage. Neighborhood tree canopy goals are calculated based on natural biome baselines (e.g., forest: 40%, grassland: 20%, desert: 15%) adjusted by building density, which limits spaces where trees can be planted. Tree canopy cover is obtained from a pre-aggregated high-resolution tree canopy dataset provided by Google Environmental Insights Explorer. The percentage of tree canopy cover is calculated as follows:\u003cdiv id=\"Equa\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equa\" name=\"EquationSource\"\u003e\n$$\\:Canopy\\:Cover\\:\\left(\\%\\right)=\\raisebox{1ex}{${A}_{T}$}\\!\\left/\\:\\!\\raisebox{-1ex}{${A}_{L},$}\\right.\\times\\:100$$\u003c/div\u003e\u003c/div\u003e,\u003c/p\u003e \u003cp\u003ewhere: \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{A}_{L}\\)\u003c/span\u003e\u003c/span\u003e is the land area of the block group, not including the water area, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{A}_{T}\\)\u003c/span\u003e\u003c/span\u003e is the tree canopy cover in the area. Building density adjustments utilize a trend line between density and canopy for each biome, with forests and Mediterranean areas receiving the most significant adjustments. The canopy gap (GAP) measures the difference between the existing tree canopy and the adjusted goal. The TES is calculated by multiplying the gap score with an index, \u003cem\u003eE\u003c/em\u003e, for prioritizing neighborhoods with the greatest need for tree planting:\u003cdiv id=\"Equb\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equb\" name=\"EquationSource\"\u003e\n$$\\:TES=100\\left(1-{GAP}_{score}\\times\\:E\\right)$$\u003c/div\u003e\u003c/div\u003e.\u003c/p\u003e \u003cp\u003eScores range from 0 to 100, with higher scores reflecting greater equity. Prior research defines Moderate Tree Equity as 70\u0026ndash;90% of tree canopy cover; High Tree Equity is more than 90%. Park need was categorized based on the following thresholds: Very High Park Need: Park need greater than 3.5; High Park Need: Park need between 3 and 3.5; Moderate Park Need: Park need between 2 and 3; Low Park Need: Park need less than 2.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eHeat extremity\u003c/strong\u003e \u003cp\u003eTPL\u0026rsquo;s urban heat severity is based on the highest extremities of summer surface temperatures averaged by block group. It is defined as the average land surface temperature of the block group in degrees Fahrenheit above or below the city's average land surface temperature (TPL Urban Heat Islands dataset 2023).\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eGreen Space within twenty-five meters of the road centerline\u003c/strong\u003e \u003cp\u003eGreen space is defined as areas containing trees and forests, shrubs, grasses, herbaceous plants, woody wetlands, and emergent wetlands, with water bodies along walkable roads provided by the EnviroAtlas (2020). The EnviroAtlas dataset measures green space within twenty-five meters of each road centerline and provides a percentage of the green space relative to the total area between street intersections.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eTree Cover within an 8.5-meter strip starting from the road edge\u003c/strong\u003e \u003cp\u003eThe percentage of tree cover containing trees, forests, and woody wetlands along walkable roads provided by the EnviroAtlas (2020). Tree cover is calculated within an 8.5-meter strip starting from the road edge (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). Low tree canopy cover was defined as less than 70% of the total tree canopy cover.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cem\u003ePark access.\u003c/em\u003e Walking access to parks was assessed using the Natural Lands Trust tool called ParkServe\u0026reg;, which identifies the percent of a community within a 10-minute walk (1/4 mile\u003c/p\u003e \u003cp\u003ebuffer of a park.\u003c/p\u003e \u003cp\u003e \u003cem\u003eHealth Vulnerability\u003c/em\u003e. Physical inactivity represents the percentage of adults (aged eighteen and older) in a census tract who report not participating in any physical activities or exercises, such as running, walking, or gardening, outside their regular job duties over the past month. Poor mental health was measured as the percentage of adults in a census block group (CBG) who reported experiencing poor mental health for 14 or more days within the past 30 days. Poor mental health includes feelings of stress, depression, or emotional challenges that affect daily functioning. These estimates are based on 2020 BRFSS data and modeled for small geographic areas.\u003c/p\u003e \u003cp\u003e \u003cem\u003eSocioeconomic Vulnerability\u003c/em\u003e. Age dependency is calculated as the ratio of non-working populations, including children (0\u0026ndash;17) and seniors (65+), to the working-age population, reflecting societal support demands. Unemployment is defined as the percentage of people in the CBG who are not working, are actively seeking employment, and are available to work. Poverty is measured as the percentage of individuals living below 200% of the federal poverty line. Linguistic isolation is the proportion of households where no individual aged fourteen or older speaks English fluently. Race and ethnicity are captured by the percentage of populations identifying as non-Latine White, including persons who self-identify as Black, American Indian, Alaska Native, Asian, Native Hawaiian, Pacific Islander, and Latine.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eAligning with the conceptual model presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, this study employs SOMs to identify patterns in environmental, social, and health burdens, such as Tree Equity Score, physical inactivity, poor mental health, and access to green space, that are associated with the risk of CPA and CN. A SOM is an artificial neural network (ANN)-based classifier that utilizes an unsupervised machine learning algorithm to cluster neighborhoods with similar characteristics, revealing hidden patterns that traditional statistical methods may overlook (Basara \u0026amp; Yuan, 2008; Kohonen, 1990). To build the SOM, a predefined surface is initialized where each neuron (or unit) is assigned a weight vector corresponding to the dimensionality of the input features. During training, the Best Matching Unit (BMU)\u0026mdash;the neuron with the smallest distance to an input vector\u0026mdash;is identified, and both the BMU and its neighboring neurons adjust their weight vectors toward the input data (Augustijn \u0026amp; Zurita-Milla, 2013; Basara \u0026amp; Yuan, 2008). While the SOM does not directly incorporate geographic coordinates, it preserves spatial coherence by grouping feature-similar and spatially proximate neighborhoods into adjacent regions of the map. This ensures the final SOM clusters retain meaningful geographic and contextual relationships when projected back into physical space (Andrienko et al., 2010). As a result, spatially closer neighborhoods with similar characteristics are more likely to be grouped, forming geographically meaningful clusters when visualized on a map.\u003c/p\u003e \u003cp\u003eFor this analysis, fifteen variables were normalized using Z-scores, and multicollinearity was assessed before training the SOM. To train the SOM, we used a 20 \u0026times; 10 hexagonal grid, with each node representing a distinct region of the feature space. The training consisted of an initial exploratory phase followed by an extended iteration phase (rlen\u0026thinsp;=\u0026thinsp;1000) to refine node weight vectors and minimize the mean distance between node weights and their assigned observations. As training progressed, data points were mapped to their BMUs, and nearby neurons adjusted their weights, preserving topological relationships. Over time, the SOM converged to a structured 2D representation of the original high-dimensional feature space. We used \u003cem\u003ek\u003c/em\u003e-means clustering to determine the optimal number of clusters based on the within-cluster sum of squares (WSS). These clusters were visualized with boundaries to highlight their spatial structure. We examined the loss function plot to evaluate convergence, and changes in node distances were visualized to identify potential cluster boundaries. Additionally, component plane plots were used to interpret the SOM by mapping key variables, such as park priority, unemployment rate, and tree cover, onto the SOM nodes, helping to reveal underlying patterns in environmental and social risk factors associated with CPA and CN. After training, the SOM nodal assignments were merged with the incident-level data to categorize each incident by its SOM cluster.\u003c/p\u003e \u003cp\u003eWe then used a negative binomial regression model to assess the relationship between SOM clusters and rates of CPA and CN. The dependent variables in the model were the number of CPA and CN cases recorded within each CBG. Given that the outcome variables are counts and the counts were overdispersed within CBGs, a negative binomial model was chosen to better account for the relatively higher variance-to-mean ratio (CN\u0026thinsp;=\u0026thinsp;1.23, CPA\u0026thinsp;=\u0026thinsp;1.60). The model included the SOM-derived cluster membership as the sole independent variable, categorizing each CBG based on its profile across multiple social, environmental, and health dimensions.\u003c/p\u003e \u003cp\u003eTo account for variations in population size across geographic units, the share of children under eighteen was included as an offset variable, enabling the model to estimate rate ratios rather than raw counts. The model was structured as follows:\u003cdiv id=\"Equc\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equc\" name=\"EquationSource\"\u003e\n$$\\:{Y}_{i}=NB({\\mu\\:}_{i},\\theta\\:)$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{Y}_{i}\\:\\)\u003c/span\u003e\u003c/span\u003eis the count of CN or CPA, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\mu\\:}_{i}\\)\u003c/span\u003e\u003c/span\u003e is the expected count, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\theta\\:\\)\u003c/span\u003e\u003c/span\u003eis the overdispersion parameter. The log-link function applied to the data is given by:\u003cdiv id=\"Equd\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equd\" name=\"EquationSource\"\u003e\n$$\\:\\text{log}\\left(E\\left({Y}_{i}\\right)\\right)={\\beta\\:}_{0}+{\\beta\\:}_{1}{SOM}_{i}+\\text{l}\\text{o}\\text{g}\\left(pop\\right)$$\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003ewhere log(\u003cem\u003epop\u003c/em\u003e) is the offset term used to normalize the CN and CPA counts within each CBG, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\beta\\:}_{1}\\)\u003c/span\u003e\u003c/span\u003eis the effect of SOM cluster membership on the rate of CN and CPA.\u003c/p\u003e \u003cp\u003eTo quantify differences between clusters, rate ratios (RRs) were computed along with 95% confidence intervals (CIs). A rate ratio greater than one indicates a higher rate of neglect or physical abuse in the comparison cluster relative to the reference cluster. To reduce the risk of Type I error, we applied a correction for multiple comparisons, ensuring that statistically significant results remained robust. The regression model was conducted using the MASS package's \u003cem\u003eglm.nb\u003c/em\u003e function in R v 4.4.2.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003e \u003cem\u003eDescriptive Results.\u003c/em\u003e The correlation analysis reveals significant relationships between tree equity, park priority, green space, and socioeconomic indicators (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Tree equity is positively correlated with green space (%) (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.69, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and tree cover (%) (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.73, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), while negatively associated with the percent of non-White persons (\u003cem\u003er\u003c/em\u003e = -0.76, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and park need (\u003cem\u003er\u003c/em\u003e = -0.67, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Park priority is positively correlated with the percentage of poverty (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.69, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001), greater physical inactivity (r\u0026thinsp;=\u0026thinsp;0.66, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and higher mental health concerns (r\u0026thinsp;=\u0026thinsp;0.53, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Other notable correlations include low physical activity and poverty (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.55, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and mental health issues and poverty (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.44, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Green space has a negative correlation with park need (\u003cem\u003er\u003c/em\u003e = -0.62, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWhile traditional correlation analysis identifies pairwise relationships, the component planes plot (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e) reveals the co-occurrence of social, physical, and environmental characteristics across geographic areas. As expected, TES, tree cover, and green space exhibit similar distributions, indicating that areas with higher tree coverage and green space also have higher tree equity scores. In contrast, an inverse correlation between park need and park priority reflects the increased park need in areas dominated by historically marginalized groups. The share of the non-White population exhibits a similar spatial clustering pattern to unemployment and poverty, with higher values concentrated in the same regions. Poor mental health and low physical activity also overlap, suggesting that areas with poor mental health tend to have lower physical activity levels. Temperature anomaly aligns with areas of lower tree cover and green space, indicating that regions with less green infrastructure experience more significant temperature fluctuations.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe component planes plot also reveals the interrelationship between environmental, physical, and social factors. For example, low TES overlap with areas with high intersection density, high poverty rates, high unemployment, and a greater proportion of non-White residents. Regions with greater intersection density also tend to align with higher poverty and unemployment rates, as well as a greater percentage of non-White residents. In contrast, areas with higher TES, which indicate better access to urban tree coverage, tend to have lower intersection density, poverty, and unemployment rates. These areas also have fewer non-White residents, highlighting disparities in the distribution of tree canopies and access to green infrastructure across different social and economic groups.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDescriptive characteristics of socio-environmental characteristics within Self-Organizing Map clusters\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCluster 1: Urban, High Vulnerability, Low Park, and Tree Equity\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;84; 3.2%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCluster 2: Urban, High Vulnerability, High Park\u003c/p\u003e \u003cp\u003eNeed, Moderate Tree Equity\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;225; 8.6%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCluster 3: Suburban, Low Vulnerability, Moderate Park Need, High\u003c/p\u003e \u003cp\u003eTree Equity\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;544; 20.9%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eCluster 4: Urban, Mixed Vulnerability, High Park Need, Moderate\u003c/p\u003e \u003cp\u003eTree Equity\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;385; 14.8%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eCluster 5:\u003c/p\u003e \u003cp\u003eSuburban, Moderate Vulnerability, Moderate Park Need, High\u003c/p\u003e \u003cp\u003eTree Equity\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;391; 15.0%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eCluster 6: Suburban, Low Vulnerability, Low Park Need, High Tree Equity\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;470; 18.1%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eCluster 7: Suburban, Lowest Vulnerability, Lowest Park Need, Highest Tree\u003c/p\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;504; 19.4%)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTree Equity Score\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e64.78 (8.09)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e68.98 (8.26)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e91.18 (8.19)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e82.93 (8.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e74.06 (9.16)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e86.30 (9.00)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e77.67 (6.68)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePark Priority\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3.69 (0.53)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3.50 (0.53)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.89 (0.61)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.63 (0.57)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e3.07 (0.50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e2.10 (0.41)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e2.66 (0.50)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGreen Space\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e20.37 (7.23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e24.02 (7.18)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e40.17 (10.57)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e26.44 (9.58)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e25.12 (8.41)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e35.52 (7.48)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e29.35 (5.53)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTree Canopy\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e19.90 (7.80)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e21.88 (6.28)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e36.50 (11.50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e28.58 (10.37)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e22.92 (8.17)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e31.34 (8.69)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e28.93 (6.13)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePark Need\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3.69 (0.53)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e3.50 (0.53)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.89 (0.61)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.63 (0.57)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e3.07 (0.50)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e2.10 (0.41)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e2.66 (0.50)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePark Access\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.23 (0.31)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.30 (0.27)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.34 (0.31)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.27 (0.34)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.28 (0.27)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.50 (0.26)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.53 (0.30)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIntersection Density\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e59.01 (13.36)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e65.32 (11.61)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e62.51 (24.40)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e65.51 (13.69)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e65.03 (10.73)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e56.46 (12.79)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e53.29 (10.36)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePop/Acre\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e42.52 (21.04)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e38.92 (19.45)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e20.47 (17.35)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e40.83 (23.24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e37.11 (20.70)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e16.14 (7.78)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e22.11 (11.29)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUnemployment\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.09 (0.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.09 (0.06)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.06 (0.06)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.09 (0.08)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.07 (0.05)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.07 (0.06)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.06 (0.05)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDependency Ratio\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.58 (0.19)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.52 (0.19)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.60 (0.37)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.28 (0.24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.49 (0.25)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.58 (0.23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.53 (0.21)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLow Physical Activity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e28.09 (7.67)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e25.23 (7.49)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e11.70 (4.91)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e5.24 (6.68)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e21.15 (7.90)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e15.03 (5.54)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e18.85 (8.05)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePerson of Color\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.93 (0.12)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.90 (0.14)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.39 (0.20)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.50 (0.23)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.77 (0.21)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.53 (0.25)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.74 (0.21)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePoor Mental Health\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e18.00 (4.55)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e16.44 (4.28)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e10.76 (3.73)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e5.30 (6.40)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e15.08 (4.37)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e12.20 (3.55)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e13.46 (5.26)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePoverty\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.55 (0.16)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.47 (0.16)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.16 (0.12)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.31 (0.19)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.40 (0.20)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.19 (0.12)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.31 (0.15)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eChildren\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.27 (0.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.22 (0.09)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.16 (0.09)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.09 (0.08)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.19 (0.10)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.18 (0.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.20 (0.08)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThe Self-Organizing Map (SOM) analysis identified distinct neighborhood clusters based on environmental, social, and health indicators. Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e presents the mean values of each variable within the clusters. Figure\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e maps the SOM clusters across the city.\u003c/p\u003e \u003cp\u003e \u003cem\u003eCluster 1: Urban, High Vulnerability, Low Park, and Tree Equity (N\u0026thinsp;=\u0026thinsp;84; 3.2%).\u003c/em\u003e Cluster 1 represents the highest social, health, and environmental vulnerability levels in densely populated urban areas. These neighborhoods have low green space coverage (20.37%), limited tree canopy (19.90%), and the lowest tree equity score (64.78). Park need is very high (3.69), and only 23% of residents can access a park within a 10-minute walk. The built environment is dense, with high intersection density (59.01) and a population density of 42.52 people per acre. Socioeconomic conditions reflect severe disadvantage, with relatively high unemployment (9%), elevated poverty (55%), and a high dependency ratio (0.58). The population is predominantly people of color (93%), with low physical activity levels (28.09%) and the highest prevalence of poor mental health (18.00%). Children make up 27% of the population.\u003c/p\u003e \u003cp\u003e \u003cem\u003eCluster 2: Urban, High Vulnerability, High Park Need, Moderate Tree Equity (N\u0026thinsp;=\u0026thinsp;225; 8.6%).\u003c/em\u003e Cluster 2 shares many characteristics with Cluster 1, including high social and health vulnerability, but has slightly better environmental conditions. Tree equity (68.98), green space (24.02%), and tree canopy (21.88%) are somewhat improved. Park need remains high (3.50), with 30% of residents living more than a 10-minute walk from a park. The area is highly urbanized, with an intersection density of 65.32 and a population density of 38.92 people per acre. Unemployment (9%), poverty (47%), and the percentage of residents of color (90%) remain high. Physical activity is low (25.23%), and poor mental health is prevalent (16.44%). Children comprise 22% of the population.\u003c/p\u003e \u003cp\u003e \u003cem\u003eCluster 3: Suburban, Low Vulnerability, Moderate Park Need, High Tree Equity (N\u0026thinsp;=\u0026thinsp;544; 20.9%).\u003c/em\u003e Cluster 3 reflects low social vulnerability and moderate health and environmental vulnerability levels in suburban areas. Tree equity is high (91.18%), with green space (40.17%) and tree canopy coverage (36.50%) among the highest. Park need is low (1.89), although 34% of residents lack access to a park within a 10-minute walk. The built environment is moderately dense, with an intersection density of 62.51 and a population density of 20.47 per acre. Unemployment (6%) and poverty (16%) are low, and the population includes fewer people of color (39%). Physical activity levels are higher, and poor mental health (10.76%) is lower than in more vulnerable clusters. Children make up 16% of the population.\u003c/p\u003e \u003cp\u003e \u003cem\u003eCluster 4: Urban, Mixed Vulnerability, High Park Need, Moderate Tree Equity (N\u0026thinsp;=\u0026thinsp;385; 14.8%).\u003c/em\u003e Cluster 4 has mixed levels of vulnerability, combining urban density (40.83 people per acre) with moderate environmental conditions: tree equity (82.93), green space (26.44%), and tree canopy (28.58%). Park need remains high (2.63), and only 27% of residents live within a 10-minute walk of a park. The population is 50% people of color. Despite having a high intersection density (65.51), this cluster has the lowest levels of physical inactivity (5.24%) and the lowest poor mental health prevalence (5.30%). However, social conditions vary: poverty is moderate (31%), unemployment is high (9%), and the dependency ratio is low (0.28). Children make up 9% of the population.\u003c/p\u003e \u003cp\u003e \u003cem\u003eCluster 5: Suburban, Moderate Vulnerability, Moderate Park Need, High Tree Equity (N\u0026thinsp;=\u0026thinsp;391; 15.0%).\u003c/em\u003e Cluster 5 reflects moderate social, health, and environmental vulnerability. Tree equity (74.06), green space (25.12%), and tree canopy (22.92%) are moderate. Park need is elevated (3.07), and only 28% of residents have nearby park access. The area is moderately dense (37.11 people per acre; intersection density 65.03). Poverty (40%) and unemployment (7%) suggest moderate disadvantage. Residents are 77% people of color, and poor mental health (15.08%) and physical inactivity (21.15%) are relatively high. Children make up 19% of the population.\u003c/p\u003e \u003cp\u003e \u003cem\u003eCluster 6: Suburban, Low Park Need, Low Social Vulnerability, Moderate Health Burden (N\u0026thinsp;=\u0026thinsp;470; 18.1%).\u003c/em\u003e Cluster 6 demonstrates low social and environmental vulnerability, with some moderate health concerns. Tree equity (86.30), green space (35.52%), and tree canopy (31.34%) are high. Park need is low (2.10), and 50% of residents have park access, which is one of the highest levels. It is a lower-density area (16.14 people per acre; intersection density 56.46). Socially, this cluster shows low poverty (19%), unemployment (7%), and a moderate racial composition (53% people of color). Physical inactivity (15.03%) and poor mental health (12.20%) suggest relatively lower levels of physical activity and poorer mental health compared to clusters 3 and 4, but higher than clusters 1 and 2. Children make up 18% of the population.\u003c/p\u003e \u003cp\u003e \u003cem\u003eCluster 7: Suburban, Lowest Vulnerability, Lowest Park Need, Highest Tree Equity (N\u0026thinsp;=\u0026thinsp;504; 19.4%).\u003c/em\u003e Cluster 7 represents the least vulnerable cluster overall. It has the highest tree equity (94.98%), the most green space (46.10%), and the highest tree canopy (41.25%). Park need is lowest (1.43), and 53% of residents have park access, the highest across clusters. This cluster is also the least dense, with a population density of 11.71 people per acre and the lowest intersection density (53.29). It has the lowest poverty (12%), lowest unemployment (6%), and lowest percentage of people of color (28%). Physical inactivity (10.70%) and poor mental health (9.97%) are low. Children comprise 20% of the population.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003esimplifies the clusters by social, health and environmental vulnerability.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCluster\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSocial Vulnerability\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHealth Vulnerability\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eEnvironmental Vulnerability\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ePopulation\u003c/p\u003e \u003cp\u003eDensity\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eCluster Summary\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCluster 1: Urban, High Vulnerability, Low Park, and Tree Equity (N\u0026thinsp;=\u0026thinsp;84; 3.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e★★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e★★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e★★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e★★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eHighest social disadvantage, poor health outcomes, degraded environmental conditions, and extreme urban density.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCluster 2: Urban, High Vulnerability, High Park Need, Moderate Tree Equity (N\u0026thinsp;=\u0026thinsp;225; 8.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e★★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e★★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e★★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eHigh social and health vulnerability with slightly better environmental features than Cluster 1, but still dense and disadvantaged.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCluster 3: Suburban, Low Vulnerability, Moderate Park Need, High Tree Equity (N\u0026thinsp;=\u0026thinsp;544; 20.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eLow social vulnerability and moderate health burden; strong tree equity but limited park access and moderate density\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCluster 4: Urban, Mixed Vulnerability, High Park Need, Moderate Tree Equity (N\u0026thinsp;=\u0026thinsp;385; 14.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"No\" id=\"Taba\" border=\"1\"\u003e \u003ccolgroup cols=\"1\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModerate levels of social and health vulnerability, with persistent park need and urban form.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCluster 5: Suburban, Moderate Vulnerability, Moderate Park Need, High Tree Equity (N\u0026thinsp;=\u0026thinsp;391; 15%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eModerate social and health burden; good tree equity but barriers to park access remain in a moderately dense setting\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCluster 6: Suburban, Low Vulnerability, Low Park Need, High Tree Equity (N\u0026thinsp;=\u0026thinsp;470; 18.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e★★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eLow social vulnerability and low population density; health and environmental conditions are favorable but not optimal.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCluster 7: Suburban, Lowest Vulnerability, Lowest Park Need, Highest Tree (N\u0026thinsp;=\u0026thinsp;504; 19.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e★\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eLowest vulnerability across all domains\u0026mdash;social, health, environmental, and density\u0026mdash;serves as the reference group\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"6\" nameend=\"c6\" namest=\"c1\"\u003e \u003cp\u003e\u003cem\u003eNotes.\u003c/em\u003e Summary of the clusters with three different versions of stars to indicate high (★★★), moderate (★★), or low (★) for each category: social, health, environmental, and density. \u003cb\u003eSocial vulnerability\u003c/b\u003e: Based on the average of people of color, unemployment, and poverty rates. \u003cb\u003eHealth\u003c/b\u003e: Based on the average of low physical activity and poor mental health rates. \u003cb\u003eEnvironmental\u003c/b\u003e: Based on the average of tree equity, green space, and tree canopy, park need, proximity to green space, temperature anomalies, and intersection density. \u003cb\u003eDensity\u003c/b\u003e: Based on the population per acre.\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe SOM clustering visualization (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e) reveals important patterns in socio-environmental conditions across neighborhoods. The legend shows the multiple variables incorporated into this analysis, ranging from environmental factors (TES, park access, extreme heat) to socioeconomic indicators (unemployment, poverty) and demographic characteristics (dependency ratio, non-white population). The hexagonal grid organizes high-dimensional data into a two-dimensional representation where similar neighborhoods are positioned closer together. The black boundary lines between clusters highlight transition zones where neighborhood characteristics change significantly. These boundaries indicate areas of sharp differentiation in the underlying variables, including park access, extreme heat exposure, population density, socioeconomic factors, and health metrics. Clusters 3, 5, 6, and 7 occupy larger portions of the SOM, indicating these neighborhood types are more prevalent throughout the city. Cluster 2 (yellow) appears centrally located and tightly packed, suggesting a distinct neighborhood typology with consistent characteristics. Clusters 5 (blue) and 6 (purple) are adjacent on the map, indicating a gradual transition between their environmental and social characteristics. Similarly, Cluster 3 (green) and Cluster 7 (pink/magenta) occupy the right portion of the map, showing they share certain similarities while maintaining distinct profiles.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e\n\u003ch3\u003eNegative Binomial Regression Results for Child Neglect and Physical Abuse\u003c/h3\u003e\n\u003cp\u003eThe negative binomial regression model indicated statistically significant differences in CN and CPA rates across the SOM-defined neighborhood clusters (CN: Omnibus test \u003cem\u003eX\u003c/em\u003e\u0026sup2; = 15.084, df\u0026thinsp;=\u0026thinsp;5, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.020; CPA: \u003cem\u003eX\u003c/em\u003e\u0026sup2; = 26.103, df\u0026thinsp;=\u0026thinsp;5, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e \u003cp\u003eFor CN, neighborhoods in Cluster 1 (Urban, High Vulnerability, Low Park, and Tree Equity) had the highest rates of child neglect. Compared to Cluster 1, CN rates were significantly lower in Cluster 4 (Urban, Mixed Vulnerability, High Park Need, Moderate Tree Equity) (IRR\u0026thinsp;=\u0026thinsp;0.54, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) and Cluster 7 (Suburban, Lowest Vulnerability, Lowest Park Need, Highest Tree Equity) (IRR\u0026thinsp;=\u0026thinsp;0.46, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), indicating a strong gradient in risk aligned with social and environmental disadvantage. A similar pattern emerged for CPA: rates were highest in Cluster 1, with CPA incidence 2.19 times higher than in Cluster 7 (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and 1.97 times higher than in Cluster 5 (Suburban, Moderate Vulnerability, Moderate Park Need, High Tree Equity) (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001). In contrast, Cluster 3 (Suburban, Low Vulnerability, Moderate Park Need, High Tree Equity) had significantly lower CPA rates than Cluster 1 (IRR\u0026thinsp;=\u0026thinsp;0.27, p\u0026thinsp;\u0026lt;\u0026thinsp;0.05), further supporting the association between socio-environmental advantage and reduced maltreatment risk.\u003c/p\u003e \u003cp\u003eTo explore how these patterns relate to structural disadvantage, we plotted CN and CPA rates against poverty and racial composition across neighborhoods, grouped by SOM cluster. The scatterplots in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e help visualize how rates of CN and CPA vary jointly with poverty levels and the percentage of non-White residents within each cluster. As shown by the figure, higher CN and CPA rates are concentrated across both panels in the upper-right quadrant, corresponding to neighborhoods with relatively higher poverty levels and proportions of non-White residents. This trend is particularly evident in clusters characterized by high social and environmental vulnerability, where maltreatment rates are not only elevated but also more densely concentrated. This visualization complements the results of the negative binomial regression by illustrating that CN and CPA are shaped by the combined burden of racialized economic disadvantage and socioeconomic disadvantage, particularly in urban contexts with limited access to green space and other protective community resources including accessible parks, tree canopy coverage, walkable environments, and protections against extreme heat.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study applies a socio-ecological lens rooted in environmental justice to highlight the significance of environmental assets in promoting physical transformation and enhancing social well-being. Employing SOMs within the context of environmental justice, this research uncovers patterns among different environmental, social, and health variables drawn from a range of georeferenced datasets. Through the integration of spatial distances and hierarchical clustering, the analysis revealed seven unique, geographically cohesive clusters that showcase diverse neighborhood conditions. These clusters were then used to predict CN and CPA risk. Incorporating spatial proximity ensured that the clusters were both data-driven and spatially meaningful, allowing for a more accurate assessment of the geographic distribution of risk factors. The findings revealed significant correlations between low equity in parks and trees within urban areas and higher rates of CPA and CN. These results are consistent with Green (2022), who underscores the multiscalar nature of child maltreatment risk, where neighborhood-level environmental conditions, such as green space access and urban density, intersect with broader structural and social disadvantage to shape child welfare outcomes (Green, 2022). Together, these findings highlight the need for place-based, equity-focused prevention strategies that address the physical and social dimensions of risk across geodemographic groups.\u003c/p\u003e \u003cp\u003ePrevious research examining physical and social neighborhood characteristics on CN and CPA rates has yielded mixed results. While prior work found that certain physical conditions (e.g., abandoned dwellings, physical barriers) were associated with reduced maltreatment, other features, such as better-maintained streets, residential decorations, and adequate public transit, were protective for child maltreatment (McDonell \u0026amp; Skosireva, 2009). These divergent findings may reflect limitations of traditional modeling approaches that assume linear and independent relationships between predictors and outcomes. In contrast, the present study used SOMs to capture complex, nonlinear interactions among environmental, social, and health-related indicators of risk. Rather than evaluating variables independently, the SOM approach grouped neighborhoods into clusters based on multidimensional similarity across multiple indicators, producing a typology that more accurately reflects the interaction between socio-environmental vulnerabilities.\u003c/p\u003e \u003cp\u003eOur findings align with prior research demonstrating an inverse relationship between green space equity, measured through park accessibility, tree canopy coverage, and equity scores, and socioeconomic vulnerability (Cheruvalath et al., 2022; Liu et al., 2021). Previous studies have similarly found that neighborhoods with higher poverty levels, unemployment, and racialized disadvantage are more likely to suffer from poor environmental quality and reduced access to green infrastructure (Heo et al., 2021). Families living in these under-resourced areas often experience heightened parenting stress due to chronic exposure to both environmental hazards and structural disadvantage (Balseviciene et al., 2014). These stressors\u0026mdash;compounded by limited access to green space, urban heat exposure, and high traffic or intersection density\u0026mdash;can exacerbate physical and mental health challenges, diminishing caregiver capacity and increasing the risk of child maltreatment (Astell-Burt \u0026amp; Feng, 2019; Balseviciene et al., 2014). Greening interventions in high-risk communities can reduce these burdens by promoting emotional regulation, enhancing physical activity, fostering positive social interactions, and improving parental well-being. However, anti-poverty strategies alone are insufficient to address the broad range of risk factors associated with CPA and CN. Policies must also address the unequal distribution of environmental resources within socioeconomically vulnerable areas. Without explicit attention to disparities in the built environment, structural inequalities will likely persist or worsen, limiting the effectiveness of otherwise well-intentioned public health interventions.\u003c/p\u003e \u003cp\u003eAn environmental justice approach prioritizes equity in the distribution of environmental resources while addressing the intersectional impacts of social and environmental disadvantage. Effective interventions must be comprehensive, addressing individual, family-level, and structural barriers such as insufficient green infrastructure and uneven urban development that have disproportionately affected communities of color. Targeted investments in historically underserved neighborhoods are essential for tackling structural injustices and promoting the well-being of children and their families. One example of such an initiative is the Green Community Mapping Project in Kalamazoo, which aims to identify high-priority areas for expanding access to green spaces (The Trust for Public, n.d.; Tuddenham, n.d.). Through a partnership among the Children and Nature Network, Trust for Public Land, and the Kellogg Foundation, the project has identified socioeconomically vulnerable neighborhoods where children lack access to green spaces within a ten-minute walk. The project\u0026rsquo;s findings have been used to guide strategic urban planning, enabling city agencies and community organizations to embed green space development within broader health promotion strategies. Although the project primarily targets physical health outcomes, such as reducing obesity and improving overall well-being, our study highlights its relevance for violence prevention. Specifically, our findings demonstrated that the highest rates of CN and CPA were concentrated in neighborhoods with high social disadvantage and environmental deprivation\u0026mdash;areas where access to green space and tree equity were the lowest. Initiatives like the Green Community Mapping Project are consistent with our results in prioritizing place-based interventions in communities where intersecting vulnerabilities are greatest. By expanding green space access in high-risk areas, such efforts have the potential to reduce parenting stress, improve mental health, and strengthen community networks to mitigate child maltreatment risk.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eLimitations\u003c/h2\u003e \u003cp\u003eThis study contributes new insights to the literature on child welfare by utilizing SOMs as a means to capture better geospatial influences on the intersection of neighborhood and individual factors that increase the risk of child maltreatment. However, several limitations should be considered when interpreting these findings. First, we analyze this data within specific geographic areas, particularly across Los Angeles, which is uniquely situated in terms of its landscape, resources, and political environment. Therefore, the findings from this study may not be applicable for predicting risk factors related to child maltreatment throughout the United States. Similarly, Los Angeles has a distinct climate that affects its ability to generate green space, which may be limited in other regions of California. Another limitation is that the outcome is not defined by residential location but by the event's location; thus, there is an assumption that the event occurred in the residential neighborhood, which we cannot confirm with certainty. Additionally, this study found that the most prominent cluster included neighborhoods with higher tree density and green space, alongside elevated levels of employment, low levels of poverty, high rates of physical activity, and low incidences of poor mental health; however, we cannot establish a causal link between green space and the presence of trees and these individual resilience factors based on this cross-sectional analysis. To better model these complex interactions, future research should continue to disentangle how intersecting neighborhood conditions structure child maltreatment risk, for example, using higher-order random effects models such as Multilevel Analysis of Individual Heterogeneity and Discriminatory Accuracy (MAIHDA) (C. R. Evans et al., 2024). These findings highlight the need to contextualize neighborhood features within broader ecological systems and support integrated, place-based interventions that move beyond surface-level infrastructure improvements or narrowly focused anti-poverty strategies.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eBy incorporating green space equity, defined as the equitable distribution of green resources in the most socially vulnerable neighborhoods, this study advances our understanding of how environmental justice can mitigate CN and CPA risk. Targeting built environment interventions in communities with limited access to parks, green spaces, and recreational resources is essential for protecting children and promoting safety.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eAs this study involved the secondary analysis of publicly available, de-identified data, it was not considered as human subjects research, as defined by federal regulations, and did not require approval from an ethics committee or Institutional Review Board (IRB).\u003c/p\u003e\u003ch2\u003eFunding:\u003c/h2\u003e \u003cp\u003eThe authors have no financial or proprietary interests in any material discussed in this article.\u003c/p\u003e \u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eGEBS led the conceptualization, methodology design, and data curation for the study. GEBS also wrote the main text, supervised the project, and contributed to editing and revision. SD contributed to the conceptualization of the study and writing of the manuscript. BR was involved in conceptualization, writing, and editing. CS contributed to writing and editing the manuscript. All authors reviewed and approved the final version of the manuscript.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eAll data used in this study are publicly available from open-access sources. The data can be accessed through the following links: Child Physical Abuse and Neglect: City of Los Angeles Open Data Portal, Crime Data from 2020 to Presenthttps://data.lacity.org/Public-Safety/Crime-Data-from-2020-to-Present/2nrs-mtv8. Tree Equity Score is available from American Forests at https://www.americanforests.org/initiatives/tree-equity/. The heat extremity \u0026amp; Park Equity Score can be downloaded from the Trust for Public Land\u0026rsquo;s ParkServe Database at https://parkserve.tpl.org. The U.S. Environmental Protection Agency, EnviroAtlas contains the data on green space along walkable road networks and can be downloaded at https://www.epa.gov/enviroatlas/enviroatlas-data. Physical inactivity and poor mental health was curated from the CDC PLACES Project (https://www.cdc.gov/places/) and is also available from the American Forests database. The socioeconomic indicators can be downloaded from the U.S. Census Bureau, American Community Surveyhttps://www.census.gov/programs-surveys/acs. All datasets are openly accessible, and no special access or permissions are required.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAndrienko, G., Andrienko, N., Bremm, S., Schreck, T., Von Landesberger, T., Bak, P., \u0026amp; Keim, D. (2010). Space‐in‐Time and Time‐in‐Space Self‐Organizing Maps for Exploring Spatiotemporal Patterns. \u003cem\u003eComputer Graphics Forum\u003c/em\u003e, \u003cem\u003e29\u003c/em\u003e(3), 913\u0026ndash;922. https://doi.org/10.1111/j.1467-8659.2009.01664.x\u003c/li\u003e\n\u003cli\u003eAstell-Burt, T., \u0026amp; Feng, X. (2019). Association of Urban Green Space With Mental Health and General Health Among Adults in Australia. \u003cem\u003eJAMA Network Open\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(7), e198209. https://doi.org/10.1001/jamanetworkopen.2019.8209\u003c/li\u003e\n\u003cli\u003eAugustijn, E.-W., \u0026amp; Zurita-Milla, R. (2013). Self-organizing maps as an approach to exploring spatiotemporal diffusion patterns. \u003cem\u003eInternational Journal of Health Geographics\u003c/em\u003e, \u003cem\u003e12\u003c/em\u003e(1), 60. https://doi.org/10.1186/1476-072X-12-60\u003c/li\u003e\n\u003cli\u003eAzar, S. T., \u0026amp; Weinzierl, K. M. (2005). Child Maltreatment and Childhood Injury Research: A Cognitive Behavioral Approach. \u003cem\u003eJournal of Pediatric Psychology\u003c/em\u003e, \u003cem\u003e30\u003c/em\u003e(7), 598\u0026ndash;614. https://doi.org/10.1093/jpepsy/jsi046\u003c/li\u003e\n\u003cli\u003eBalseviciene, B., Sinkariova, L., \u0026amp; Andrusaityte, S. (2014). Do green spaces matter? The associations between parenting stress, child mental health problems and green spaces. \u003cem\u003eProcedia-Social and Behavioral Sciences\u003c/em\u003e, \u003cem\u003e140\u003c/em\u003e, 511\u0026ndash;516.\u003c/li\u003e\n\u003cli\u003eBarboza, G., Angulski, K., Hines, L., \u0026amp; Brown, P. (2022). Variability in Opioid-Related Drug Overdoses, Social Distancing, and Area-Level Deprivation during the COVID-19 Pandemic: A Bayesian Spatiotemporal Analysis. \u003cem\u003eJournal of Urban Health\u003c/em\u003e, \u003cem\u003e99\u003c/em\u003e(5), 873\u0026ndash;886.\u003c/li\u003e\n\u003cli\u003eBarboza, G. E., \u0026amp; Schiamberg, L. (2021). Dual trajectories of parenting self‐efficacy and depressive symptoms in new, postpartum mothers and socioemotional adjustment in early childhood: A growth mixture model. \u003cem\u003eInfant Mental Health Journal\u003c/em\u003e, \u003cem\u003e42\u003c/em\u003e(5), 636\u0026ndash;654. https://doi.org/10.1002/imhj.21937\u003c/li\u003e\n\u003cli\u003eBarboza-Salerno, G. E. (2020). Cognitive readiness to parent, stability and change in postpartum parenting stress and social-emotional problems in early childhood: A second order growth curve model. \u003cem\u003eChildren and Youth Services Review\u003c/em\u003e, \u003cem\u003e113\u003c/em\u003e, 104958. https://doi.org/10.1016/j.childyouth.2020.104958\u003c/li\u003e\n\u003cli\u003eBarboza-Salerno, G. E. (2023). The neighborhood deprivation gradient and child physical abuse and neglect: A Bayesian spatial model. \u003cem\u003eChild Abuse \u0026amp; Neglect\u003c/em\u003e, \u003cem\u003e146\u003c/em\u003e, 106501.\u003c/li\u003e\n\u003cli\u003eBarboza-Salerno, G. E., \u0026amp; Meshelemiah, J. C. (2023). Gun Violence on Walkable Routes to and from School: Recommendations for Policy and Practice. \u003cem\u003eJournal of Urban Health\u003c/em\u003e, 1\u0026ndash;16.\u003c/li\u003e\n\u003cli\u003eBasara, H. G., \u0026amp; Yuan, M. (2008). Community health assessment using self-organizing maps and geographic information systems. \u003cem\u003eInternational Journal of Health Geographics\u003c/em\u003e, \u003cem\u003e7\u003c/em\u003e(1), 67. https://doi.org/10.1186/1476-072X-7-67\u003c/li\u003e\n\u003cli\u003eBywaters, P., Brady, G., Sparks, T., \u0026amp; Bos, E. (2016). Child welfare inequalities: New evidence, further questions. \u003cem\u003eChild \u0026amp; Family Social Work\u003c/em\u003e, \u003cem\u003e21\u003c/em\u003e(3), 369\u0026ndash;380. https://doi.org/10.1111/cfs.12154\u003c/li\u003e\n\u003cli\u003eCause, A. G. (2020). \u003cem\u003eThe Economically Disadvantaged Speak: Exploring the Intersection of Poverty, Race, Child Neglect and Racial Disproportionality in the Child Welfare System\u003c/em\u003e [PhD Thesis, Portland State University]. https://search.proquest.com/openview/9105306031fcc5ec3764f7575d7fbf4b/1?pq-origsite=gscholar\u0026amp;cbl=18750\u0026amp;diss=y\u0026amp;casa_token=8hpVkVmiqMIAAAAA:Nvwc3mdUlWgnBmtRZWMOl4foacEuhB2c\u003cbr\u003e__1YMnt9BdBasQvqMTrB1Sox7rgUXvD6RqfDsfrLGoM\u003c/li\u003e\n\u003cli\u003eChen, W., Wu, A. N., \u0026amp; Biljecki, F. (2021). Classification of urban morphology with deep learning: Application on urban vitality. \u003cem\u003eComputers, Environment and Urban Systems\u003c/em\u003e, \u003cem\u003e90\u003c/em\u003e, 101706. https://doi.org/10.1016/j.compenvurbsys.2021.101706\u003c/li\u003e\n\u003cli\u003eCheruvalath, H., Homa, J., Singh, M., Vilar, P., Kassam, A., \u0026amp; Rovin, R. A. (2022). Associations Between Residential Greenspace, Socioeconomic Status, and Stroke: A Matched Case-Control Study. \u003cem\u003eJournal of Patient-Centered Research and Reviews\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e(2), 89\u0026ndash;97. https://doi.org/10.17294/2330-0698.1886\u003c/li\u003e\n\u003cli\u003eChung, G., Lanier, P., \u0026amp; Wong, P. Y. J. (2022). Mediating Effects of Parental Stress on Harsh Parenting and Parent-Child Relationship during Coronavirus (COVID-19) Pandemic in Singapore. \u003cem\u003eJournal of Family Violence\u003c/em\u003e, \u003cem\u003e37\u003c/em\u003e(5), 801\u0026ndash;812. https://doi.org/10.1007/s10896-020-00200-1\u003c/li\u003e\n\u003cli\u003eElise Barboza-Salerno, G. (2024). Material Hardship, Labor Market Characteristics and Substantiated Child Maltreatment: A Bayesian Spatiotemporal Analysis. \u003cem\u003eChildren and Youth Services Review\u003c/em\u003e, \u003cem\u003e157\u003c/em\u003e, 107371. https://doi.org/10.1016/j.childyouth.2023.107371\u003c/li\u003e\n\u003cli\u003eEllis, W. E., Dumas, T. M., \u0026amp; Forbes, L. M. (2020). Physically isolated but socially connected: Psychological adjustment and stress among adolescents during the initial COVID-19 crisis. \u003cem\u003eCanadian Journal of Behavioural Science/Revue Canadienne Des Sciences Du Comportement\u003c/em\u003e, \u003cem\u003e52\u003c/em\u003e(3), 177.\u003c/li\u003e\n\u003cli\u003eEvans, C. R., Leckie, G., Subramanian, S. V., Bell, A., \u0026amp; Merlo, J. (2024). A tutorial for conducting intersectional multilevel analysis of individual heterogeneity and discriminatory accuracy (MAIHDA). \u003cem\u003eSSM - Population Health\u003c/em\u003e, \u003cem\u003e26\u003c/em\u003e, 101664. https://doi.org/10.1016/j.ssmph.2024.101664\u003c/li\u003e\n\u003cli\u003eEvans, M. F., Gazze, L., \u0026amp; Schaller, J. (2025). Temperature and maltreatment of young children. \u003cem\u003eReview of Economics and Statistics\u003c/em\u003e, 1\u0026ndash;37.\u003c/li\u003e\n\u003cli\u003eFan, Y., Das, K. V., \u0026amp; Chen, Q. (2011). Neighborhood green, social support, physical activity, and stress: Assessing the cumulative impact. \u003cem\u003eHealth \u0026amp; Place\u003c/em\u003e, \u003cem\u003e17\u003c/em\u003e(6), 1202\u0026ndash;1211.\u003c/li\u003e\n\u003cli\u003eFinno-Velasquez, M., He, A. S., Perrigo, J. L., \u0026amp; Hurlburt, M. S. (2017). Community Informant Explanations for Unusual Neighborhood Rates of Child Maltreatment Reports. \u003cem\u003eChild and Adolescent Social Work Journal\u003c/em\u003e, \u003cem\u003e34\u003c/em\u003e(3), 191\u0026ndash;204. https://doi.org/10.1007/s10560-016-0463-3\u003c/li\u003e\n\u003cli\u003eFont, S. A., \u0026amp; Maguire-Jack, K. (2015). Decision-making in child protective services: Influences at multiple levels of the social ecology. \u003cem\u003eChild Abuse \u0026amp; Neglect\u003c/em\u003e, \u003cem\u003e47\u003c/em\u003e, 70\u0026ndash;82. https://doi.org/10.1016/j.chiabu.2015.02.005\u003c/li\u003e\n\u003cli\u003eFoundation, T. A. E. C. (2021, October 12). \u003cem\u003eChild Maltreatment Trends\u003c/em\u003e. The Annie E. Casey Foundation. https://www.aecf.org/blog/child-maltreatment-trends\u003c/li\u003e\n\u003cli\u003eFreisthler, B. (2004). A spatial analysis of social disorganization, alcohol access, and rates of child maltreatment in neighborhoods. \u003cem\u003eChildren and Youth Services Review\u003c/em\u003e, \u003cem\u003e26\u003c/em\u003e(9), 803\u0026ndash;819.\u003c/li\u003e\n\u003cli\u003eFreisthler, B., Gruenewald, P. J., Treno, A. J., \u0026amp; Lee, J. (2003). Evaluating alcohol access and the alcohol environment in neighborhood areas. \u003cem\u003eAlcoholism: Clinical and Experimental Research\u003c/em\u003e, \u003cem\u003e27\u003c/em\u003e(3), 477\u0026ndash;484.\u003c/li\u003e\n\u003cli\u003eFrumkin, H., Bratman, G. N., Breslow, S. J., Cochran, B., Kahn Jr, P. H., Lawler, J. J., Levin, P. S., Tandon, P. S., Varanasi, U., Wolf, K. L., \u0026amp; Wood, S. A. (2017). Nature Contact and Human Health: A Research Agenda. \u003cem\u003eEnvironmental Health Perspectives\u003c/em\u003e, \u003cem\u003e125\u003c/em\u003e(7), 075001. https://doi.org/10.1289/EHP1663\u003c/li\u003e\n\u003cli\u003eGreen, J. W. (2022). The Built Environment and Predicting Child Maltreatment: An Application of Random Forests to Risk Terrain Modeling. \u003cem\u003eThe Professional Geographer\u003c/em\u003e, \u003cem\u003e74\u003c/em\u003e(1), 67\u0026ndash;78. https://doi.org/10.1080/00330124.2021.1970591\u003c/li\u003e\n\u003cli\u003eGuterman, N. B., Lee, S. J., Taylor, C. A., \u0026amp; Rathouz, P. J. (2009). Parental perceptions of neighborhood processes, stress, personal control, and risk for physical child abuse and neglect. \u003cem\u003eChild Abuse \u0026amp; Neglect\u003c/em\u003e, \u003cem\u003e33\u003c/em\u003e(12), 897\u0026ndash;906.\u003c/li\u003e\n\u003cli\u003eHaas, B. M., Berg, K. A., Schmidt-Sane, M. M., Korbin, J. E., \u0026amp; Spilsbury, J. C. (2018). How might neighborhood built environment influence child maltreatment? Caregiver perceptions. \u003cem\u003eSocial Science \u0026amp; Medicine\u003c/em\u003e, \u003cem\u003e214\u003c/em\u003e, 171\u0026ndash;178. https://doi.org/10.1016/j.socscimed.2018.08.033\u003c/li\u003e\n\u003cli\u003eHeo, S., Nori-Sarma, A., Kim, S., Lee, J.-T., \u0026amp; Bell, M. L. (2021). Do persons with low socioeconomic status have less access to greenspace? Application of accessibility index to urban parks in Seoul, South Korea. \u003cem\u003eEnvironmental Research Letters\u003c/em\u003e, \u003cem\u003e16\u003c/em\u003e(8), 084027.\u003c/li\u003e\n\u003cli\u003eKendrick, E. (2015). \u003cem\u003eA Neighborhood Study: Recreational Parks and Parent Stress\u003c/em\u003e [PhD Thesis, The Ohio State University]. https://kb.osu.edu/items/ffeb5391-0a3d-5081-b00f-2cf781f250b4\u003c/li\u003e\n\u003cli\u003eKohonen, T. (1990). The self-organizing map. \u003cem\u003eProceedings of the IEEE\u003c/em\u003e, \u003cem\u003e78\u003c/em\u003e(9), 1464\u0026ndash;1480.\u003c/li\u003e\n\u003cli\u003eKuo, F. E., \u0026amp; Sullivan, W. C. (2001a). Aggression and Violence in the Inner City: Effects of Environment via Mental Fatigue. \u003cem\u003eEnvironment and Behavior\u003c/em\u003e, \u003cem\u003e33\u003c/em\u003e(4), 543\u0026ndash;571. https://doi.org/10.1177/00139160121973124\u003c/li\u003e\n\u003cli\u003eKuo, F. E., \u0026amp; Sullivan, W. C. (2001b). Environment and Crime in the Inner City: Does Vegetation Reduce Crime? \u003cem\u003eEnvironment and Behavior\u003c/em\u003e, \u003cem\u003e33\u003c/em\u003e(3), 343\u0026ndash;367. https://doi.org/10.1177/0013916501333002\u003c/li\u003e\n\u003cli\u003eLe, K. (2025). The Impacts of Extreme Heat Days on the Prevalence of Domestic Abuse. \u003cem\u003eSage Open\u003c/em\u003e, \u003cem\u003e15\u003c/em\u003e(1), 21582440251317797. https://doi.org/10.1177/21582440251317797\u003c/li\u003e\n\u003cli\u003eLittleton, T., Freisthler, B., Boyd, R., Smith, A. M., \u0026amp; Barboza-Salerno, G. (2024). Historical redlining, neighborhood disadvantage, and reports of child maltreatment in a large urban county. \u003cem\u003eChild Abuse \u0026amp; Neglect\u003c/em\u003e, \u003cem\u003e156\u003c/em\u003e, 107011. https://doi.org/10.1016/j.chiabu.2024.107011\u003c/li\u003e\n\u003cli\u003eLiu, D., Kwan, M.-P., \u0026amp; Kan, Z. (2021). Analysis of urban green space accessibility and distribution inequity in the City of Chicago. \u003cem\u003eUrban Forestry \u0026amp; Urban Greening\u003c/em\u003e, \u003cem\u003e59\u003c/em\u003e, 127029.\u003c/li\u003e\n\u003cli\u003eMaguire-Jack, K., \u0026amp; Font, S. A. (2017). Community and individual risk factors for physical child abuse and child neglect: Variations by poverty status. \u003cem\u003eChild Maltreatment\u003c/em\u003e, \u003cem\u003e22\u003c/em\u003e(3), 215\u0026ndash;226.\u003c/li\u003e\n\u003cli\u003eMaguire-Jack, K., \u0026amp; Showalter, K. (2016). The protective effect of neighborhood social cohesion in child abuse and neglect. \u003cem\u003eChild Abuse \u0026amp; Neglect\u003c/em\u003e, \u003cem\u003e52\u003c/em\u003e, 29\u0026ndash;37.\u003c/li\u003e\n\u003cli\u003eMaguire-Jack, K., Yoon, S., \u0026amp; Hong, S. (2022). Social cohesion and informal social control as mediators between neighborhood poverty and child maltreatment. \u003cem\u003eChild Maltreatment\u003c/em\u003e, \u003cem\u003e27\u003c/em\u003e(3), 334\u0026ndash;343.\u003c/li\u003e\n\u003cli\u003eMcDonell, J., \u0026amp; Skosireva, A. (2009). Neighborhood Characteristics, Child Maltreatment, and Child Injuries. \u003cem\u003eChild Indicators Research\u003c/em\u003e, \u003cem\u003e2\u003c/em\u003e(2), 133\u0026ndash;153. https://doi.org/10.1007/s12187-009-9038-6\u003c/li\u003e\n\u003cli\u003eMorris, A. S., Robinson, L. R., Hays‐Grudo, J., Claussen, A. H., Hartwig, S. A., \u0026amp; Treat, A. E. (2017). Targeting Parenting in Early Childhood: A Public Health Approach to Improve Outcomes for Children Living in Poverty. \u003cem\u003eChild Development\u003c/em\u003e, \u003cem\u003e88\u003c/em\u003e(2), 388\u0026ndash;397. https://doi.org/10.1111/cdev.12743\u003c/li\u003e\n\u003cli\u003eMorton, C. M., Simmel, C., \u0026amp; Peterson, N. A. (2014). Neighborhood alcohol outlet density and rates of child abuse and neglect: Moderating effects of access to substance abuse services. \u003cem\u003eChild Abuse \u0026amp; Neglect\u003c/em\u003e, \u003cem\u003e38\u003c/em\u003e(5), 952\u0026ndash;961.\u003c/li\u003e\n\u003cli\u003eMoudon, A. V., Lee, C., Cheadle, A. D., Garvin, C., Johnson, D., Schmid, T. L., Weathers, R. D., \u0026amp; Lin, L. (2006). Operational definitions of walkable neighborhood: Theoretical and empirical insights. \u003cem\u003eJournal of Physical Activity and Health\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e(s1), S99\u0026ndash;S117.\u003c/li\u003e\n\u003cli\u003e\u003cem\u003eNew Child Maltreatment Report Finds Child Abuse and Neglect Decreased to a Five-Year Low\u003c/em\u003e. (n.d.). Retrieved December 15, 2024, from https://www.acf.hhs.gov/media/press/2023/new-child-maltreatment-report-finds-child-abuse-and-neglect-decreased-five-year\u003c/li\u003e\n\u003cli\u003e\u003cem\u003eReports of Child Abuse and Neglect, by Race/Ethnicity\u003c/em\u003e. (n.d.). Kidsdata.Org. Retrieved December 15, 2024, from https://www.kidsdata.org/topic/3/reported-abuse-race/table#fmt=1217\u0026amp;loc=2,127,347,1763,331,348,336,171,321,345,357,332,324,369,358,362,360,33\u003cbr\u003e7,327,364,356,217,353,328,354,323,352,320,339,334,365,343,330,367,344,355,366,368,265,349\u003cbr\u003e,361,4,273,59,370,326,333,322,341,338,350,342,329,325,359,351,363,340,335\u003cbr\u003e\u0026amp;tf=110\u0026amp;ch=7,11,8,10,9\u003c/li\u003e\n\u003cli\u003eRiley, T., Schleimer, J. P., \u0026amp; Jahn, J. L. (2024). Organized abandonment under racial capitalism: Measuring accountable actors of structural racism for public health research and action. \u003cem\u003eSocial Science \u0026amp; Medicine\u003c/em\u003e, \u003cem\u003e343\u003c/em\u003e, 116576. https://doi.org/10.1016/j.socscimed.2024.116576\u003c/li\u003e\n\u003cli\u003eRoof, K., \u0026amp; Oleru, N. (2008). Public Health: Seattle and King County\u0026rsquo;s Push for the Built Environment. \u003cem\u003eJournal of Environmental Health\u003c/em\u003e, \u003cem\u003e71\u003c/em\u003e(1), 24\u0026ndash;27.\u003c/li\u003e\n\u003cli\u003eThe Trust for Public. (n.d.). \u003cem\u003eKalamazoo, Michigan: Connecting Children with Nature\u003c/em\u003e. Retrieved March 17, 2025, from https://www.tpl.org/wp-content/uploads/2013/10/convis-kalamazoo.pdf\u003c/li\u003e\n\u003cli\u003eTuddenham, K. A. (n.d.). \u003cem\u003eLinking Science\u0026mdash;Measuring Health Outcomes\u003c/em\u003e. Retrieved March 17, 2025, from https://citeseerx.ist.psu.edu/document?repid=rep1\u0026amp;type=pdf\u0026amp;doi=00c244ce900cb9ed7742efe7f11a959b49c77823\u003c/li\u003e\n\u003cli\u003eU.S. Department of Health \u0026amp; Human Services, Administration for Children and Families, Administration on Children, Youth and Families, Children\u0026rsquo;s Bureau. (2024). \u003cem\u003eChild Maltreatment\u003c/em\u003e. https://www.acf.hhs.gov/sites/default/files/documents/cb/cm2022.pdf\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"child abuse and neglect, environmental justice, built environment, neighborhood risk, tree equity, green space, public health, self-organizing maps, structural inequity, Los Angeles","lastPublishedDoi":"10.21203/rs.3.rs-6521185/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6521185/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eChild abuse and neglect (CAN) represent a significant global public health challenge influenced by socioeconomic disadvantages and the built environment. While existing research has looked into the social determinants of CAN, fewer studies have focused on how environmental factors interact with social vulnerabilities to affect risk levels. This study aims to fill that gap by utilizing an environmental justice framework and Self-Organizing Maps (SOMs) to categorize neighborhoods in Los Angeles based on social, environmental, and health-related characteristics. We examined physical abuse (CPA) and child neglect (CN) from 2020 to 2023 at the census block group level. Fifteen georeferenced indicators\u0026mdash;such as poverty, tree equity, park access, heat exposure, and mental health\u0026mdash;were used as input features in a Self-Organizing Map (SOM) to identify clusters of neighborhoods with similar socio-environmental profiles. Negative binomial regression was used to predict CN and CPA rates within clusters. Seven clusters describe socio-environmental neighborhood profiles in Los Angeles. The most disadvantaged cluster was defined by high poverty rates, limited green space equity, and poor mental health, with CPA and CN rates more than double those of the most advantaged cluster. Risk levels were significantly higher in areas with intersecting social and environmental challenges. These findings highlight that structural inequities, including restricted access to green infrastructure, increase CAN risk. Our results suggest the need for targeted investments in parks, trees, and other features of the built environment in underserved neighborhoods as part of a comprehensive, place-based approach supporting healthy environments for children.\u003c/p\u003e","manuscriptTitle":"Beyond Social Disadvantage: Advancing an Environmental Justice Framework to Address Child Maltreatment Risk","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-19 06:32:49","doi":"10.21203/rs.3.rs-6521185/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"79931cbf-a771-4eaf-9d7a-c377bb8fad9d","owner":[],"postedDate":"May 19th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-04-13T16:07:11+00:00","versionOfRecord":{"articleIdentity":"rs-6521185","link":"https://doi.org/10.1007/s42448-026-00256-4","journal":{"identity":"international-journal-on-child-maltreatment-research-policy-and-practice","isVorOnly":false,"title":"International Journal on Child Maltreatment: Research, Policy and Practice"},"publishedOn":"2026-04-09 15:58:26","publishedOnDateReadable":"April 9th, 2026"},"versionCreatedAt":"2025-05-19 06:32:49","video":"","vorDoi":"10.1007/s42448-026-00256-4","vorDoiUrl":"https://doi.org/10.1007/s42448-026-00256-4","workflowStages":[]},"version":"v1","identity":"rs-6521185","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6521185","identity":"rs-6521185","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}