Does Urban Greenspace Reduce the Impact of Hot Weather on Violent Crime Rates?

This study examined whether urban greenspace modifies the relationship between ambient temperature and violent crime in Chicago, Illinois, from 2001 to 2023. Using ZIP code-level daily violent crime counts and meteorological data, researchers analyzed multiple greenspace metrics, including vegetation density, percentage of vegetated land, and tree coverage. Results showed an 8% higher violent crime risk when daily mean temperature was at the 80th percentile (25.9°C) compared to the 10th percentile (8.6°C). ZIP codes with the highest vegetation density and tree coverage had significantly lower temperature-crime associations (RR reductions of 0.036 and 0.035, respectively), suggesting greenspace can mitigate heat-related increases in violent crime.

Full Citation and Link to Article

Heo S, Choi HM, Delaney SW, James P, Bell ML. 2025. Does greenspace influence the associations between ambient temperature and violent crime? An observational study. Environmental Research Letters, 20(8), 084064. https://doi.org/10.1088/1748-9326/adef6a

Extended Summary

Central Research Question

This study investigates whether the amount and type of urban greenspace can modify the relationship between ambient temperature and violent crime. While it is well-documented that violent crime rates often increase with higher temperatures, less is known about how greenspace—through its heat mitigation and social benefits—may influence this association. The researchers focus on a single U.S. city, Chicago, Illinois, analyzing detailed neighborhood-level data to determine whether specific greenspace characteristics, such as vegetation density, percentage of vegetated land, and tree coverage, reduce the impact of elevated temperatures on violent crime rates.

Previous Literature

Prior research has consistently shown a positive relationship between higher temperatures and increased violent crime. Studies in multiple U.S. cities, such as Philadelphia, have found that violent crime risk can be several percentage points higher on the hottest days. Theoretical frameworks like the heat-aggression theory suggest that high temperatures increase discomfort and aggressive behavior, while the negative affect escape theory proposes that aggression rises at moderately high temperatures but may decline at extreme heat levels due to avoidance behaviors.

Greenspace has been widely studied for its potential benefits to public health, including physical, mental, and social well-being. Evidence indicates that it can foster social cohesion, reduce stress, and mitigate urban heat through cooling effects, particularly from tree canopies. However, the relationship between greenspace and crime is mixed. Some studies have found that vegetation is associated with lower crime rates, while others report higher crime in areas with dense or low-lying vegetation that may obscure visibility and provide cover for criminal activity. Fear of crime in parks, especially among women, has also been documented.

Importantly, previous studies have rarely examined greenspace as a modifier in the temperature-crime relationship, and most analyses have been conducted at coarse spatial scales such as cities or counties. High-resolution neighborhood-level analyses, accounting for different types of greenspace, are needed to clarify these relationships and inform urban planning decisions.

Data

The study area consisted of all 61 ZIP codes in Chicago, Illinois, covering the years 2001–2023. Violent crime data were obtained from the Chicago Police Department’s Citizen Law Enforcement Analysis and Reporting system, which includes incident dates, crime types (based on Illinois Uniform Crime Reporting codes), and locations. The violent crime category encompassed homicide, assault, battery, domestic violence, intimidation, and aggravated ritualism.

Daily meteorological data, including mean temperature and dew point temperature, were sourced from the PRISM Climate Group at 4 km resolution and aggregated to ZIP codes. Greenspace was characterized using multiple metrics:

• Enhanced Vegetation Index (EVI) from Landsat imagery for vegetation density and health

• Percentage of vegetated land from the National Land Cover Database

• Number and total area of parks from OpenStreetMap

• Percentage of recreational vegetated land (parks, cemeteries, open green areas, forests)

• Percent tree canopy coverage from NASA MODIS data

• Percent street-view tree and other vegetation coverage from Google Street View imagery analyzed with deep learning methods

Socioeconomic covariates were included from the American Community Survey (2011–2022) and the 2022 Area Deprivation Index, covering income, education, poverty, age structure, and deprivation rankings.

Methods

The analysis used a two-stage modeling approach.

Stage one applied a generalized additive model (GAM) for each ZIP code to estimate the association between daily mean temperature (lagged average over the current and previous day) and daily violent crime counts. The reference temperature was the 10th percentile (8.6°C) of warm-season (May–September) temperatures, and the target comparison was the 80th percentile (25.9°C). The models adjusted for long-term temporal trends, dew point temperature, and day of the week.

Stage two pooled the ZIP code-specific estimates using a random-effects meta-analysis. Meta-regressions were conducted to test whether greenspace metrics modified the temperature-crime relationship. Greenspace variables were grouped into tertiles, with the lowest tertile as the reference. Separate single-variable meta-regressions identified significant modifiers, and significant ones were further tested in multivariable meta-regressions adjusting for socioeconomic modifiers such as deprivation level.

Effect modification was tested using differences in log relative risks (RRs) between tertile groups, with p-values from omnibus tests and post-hoc comparisons.

Findings/Size Effects

From 2001–2023, the dataset included 1,075,959 violent crime incidents. The pooled association showed an inverted J-shaped relationship between temperature and violent crime. Violent crime risk increased steadily from cooler temperatures to a peak at around 27°C, then declined at higher temperatures. The pooled RR at the 80th percentile temperature versus the 10th percentile was 1.085 (95% CI: 1.070–1.100), corresponding to an 8% increase in violent crime risk.

Several greenspace measures significantly modified this association:

• Enhanced Vegetation Index (EVI): ZIP codes in the highest tertile had an RR of 1.085 (95% CI: 1.040–1.131) versus 1.124 (95% CI: 1.088–1.162) in the lowest tertile. This corresponds to a reduction in the temperature-crime association of about 0.039 in RR.

• Percent vegetated land: Highest tertile had RR of 1.057 (95% CI: 1.033–1.081) versus 1.100 (95% CI: 1.076–1.126) in the lowest tertile. Reduction ~0.043 in RR.

• Percent tree coverage: Highest tertile had RR of 1.088 (95% CI: 1.046–1.132) versus 1.123 (95% CI: 1.086–1.162) in the lowest tertile. Reduction ~0.035 in RR.

Number of parks, total park area, and percentage of recreational vegetated land were not significant modifiers. Street-view vegetation measures also did not significantly modify the association.

The deprivation index emerged as a significant socioeconomic modifier, with higher deprivation linked to stronger temperature-crime associations. Adjusting for deprivation did not eliminate the protective effect of vegetation density and tree coverage.

The authors interpret these results as consistent with two plausible pathways: (1) physical cooling from dense vegetation, especially trees, which can reduce heat exposure and thereby temper heat-related aggression, and (2) psychological and social benefits from greener surroundings, which may buffer stress and promote pro-social behavior.

However, the null findings for park-related measures may indicate that not all greenspaces provide the same benefits. Park quality, safety, and amenities—which were not measured—may influence their effectiveness. Similarly, street trees may not reduce overall heat exposure beyond pedestrian spaces and may have mixed thermal effects depending on local street canyon geometry.

Conclusion

The study concludes that greenspace can mitigate the positive association between high temperatures and violent crime in urban areas, but this effect depends on the type and density of vegetation. Satellite-based vegetation density, total vegetated land percentage, and tree canopy coverage emerged as consistent protective factors. The findings support incorporating greenery—particularly tall, dense vegetation like tree canopies—into urban planning as a strategy to reduce heat-related violent crime.

The authors emphasize that their work is the first high-resolution, multi-metric analysis of greenspace as a modifier of the temperature-crime link in Chicago. Strengths include the 23-year dataset, fine spatial scale (ZIP code level), and diverse greenspace metrics. Limitations include reliance on police-reported crime data, lack of data on park quality and safety, absence of detailed tree species and physical characteristics, and temporal mismatch between greenspace data (mostly from recent years) and the full crime-temperature time series.

Despite these limitations, the findings provide actionable insights for policymakers. Increasing vegetation density and tree coverage in neighborhoods—especially those with high deprivation—could serve as a crime prevention and public health strategy, particularly as climate change increases the frequency and severity of heat events. Future research should examine the mechanisms in greater detail, including the roles of greenspace quality, species selection, and social factors, and replicate the findings in other urban contexts.