Understanding the impact of urban heat islands on crime: insights from temperature, population density, and green canopy cover

Emergent literature suggests that micro-and macro-climates influence criminal behaviour; a complex phenomenon that is still incipient in theory development. This mixed-method research starts with a systematic review of the literature on the theoretical premises that urban heat islands amplify aggressive behaviour and crime. Further, it discusses the potential implications of the relationship between the environment and social outcomes on the design and planning of urban environments. A meta-synthesis was conducted to explain the correlations between patterns of criminal behaviour and thermal (dis)comfort. This correlation is to relate fundamental urban design principles to socially sustainable communities that dissuade violence and crime, and otherwise show poorly designed spaces do propagate criminal behaviour. Cross-validation was undertaken using a case study of Midland, a suburb of Perth, Western Australia. Data involves population demographics, temperature, and crime statistics relating to Offences against the Person. Analysis focused on homicide, assault, threatening behaviour, and robbery. The findings imply a positive correlation between long-term temperature, crime, population density, and green canopy cover. Variables of climate (e.g., short-to-long-term climate-related stressors) and crime types also show non-linear association. Nonetheless, forecasting the future of violence and trends of crime through attributes or potential impacts of heat and urban canopy cover on the built environment will inform sustainable social development policy, environmental planning, development strategy, designers, and planners. Recommendations are made around these in relation to making urban communities adaptive to the impacts of global warming and future densification.


Introduction
Normative studies agree on how climate change impacts various aspects of city life.Evidence in a study of the Maldives by Moosa et al. (2020) shows there is significant disparity in location-based climate change outcomes.They emphasised the implications of their study to culture, employment, food production, economic output, and mobility patterns (Su et al., 2023), posit that mobility may increase propensity for opportunistic urban crimes resulting from density, mixing, and movement of a diversity of people across public transport nodes (Brantingham & Brantingham, 1995;Felson, 2003;Madensen & Eck, 2012), with existing research finding that the impact to the surrounding increasingly urbanised environments was not inconsequential (Loukaitou-Sideris et al., 2002;Moreira & Ceccato, 2021;Newton et al., 2014).
Yet, as urbanisation continues to accelerate to meet the demands of growing populations, frequently they result in unintended and complex consequences.The impact of the constructed environment on the severity of criminal activities is a significant factor (Silva & Li, 2020) and has been the subject of increasing scholarly attention within the realm of crime place studies (Hipp et al., 2018;Sohn, 2016).For example, excessive heat in localised areas, referred to as the urban heat island effect (UHIE), 1 occurs when an urban area has significantly disparate temperatures to that of its surrounding rural belt, caused by heat absorption (Aguiar, 2012).Accordingly, it is incumbent on governments across the world to improve long-term decision making in respect of regional planning and development to mitigate unintended community harms.This must be done with cognisance of paucity of arable land and spikes in resource demand and in relation to population growth.
Moreover, giving up spaces through development and densification pressures agricultural fringes.This also affects the long-term sustainability of natural ecosystems and biodiversity (Armstrong, 2001).In addition, exposed areas required to replenish underground water supplies are impacted, as well as other environmental and socioeconomic issues.Criminal activities are influenced by demographic and socioeconomic variables (Huang et al., 2011) as well as the built environment (Fan & Sengupta, 2021;Jung et al., 2023), which serves as an external factor that impacts an individual's attitudes and behaviours (Cozens, 2011;Davies & Johnson, 2014;Eck & Weisburd, 2015).Thus, under such circumstances, maximising the potential use of all spaces to mitigate HIE, even seemingly insignificant spaces, and pedestrianising hard infrastructure towards active transportation modalities, thus reverting the vehicular-dependence paradigms which fuel current societal norms, may contribute to reducing microclimates, improving the perception of safety, and concurrently reducing the propensity for crime.
Many scholarly works have examined the role of genetics (a person's biological nature) and the environment (the natural and built environment a person exists within, including society and its institutions) in shaping or triggering criminal behaviour (Benson, 2013;Mednick & Volavka, 1980;Rafter, 2016).Most studies suggest that criminality is influenced by both factors, with the environment shaping an individual while their biological nature sets the boundaries of their behaviour (Boyanowsky, 2020;Mehta & Binder, 2012;Wilson and Herrnstein, 1998).
The identification of key factors that contribute to criminality in society has been a focus of previous research.Gaining an understanding of the factors that exert influence on criminal activities is paramount in reducing the detrimental and expensive consequences they can have on society (Twinam, 2017).These factors include population density (Watts, 1931), mental health (Fleischman et al., 2014), substance abuse (Grann & Fazel, 2004), mobility (Kelly, 2000), law enforcement (Fafchamps & Moser, 2003), public perception (Gaubatz, 1995), quality of governance (Curran, 1998), and Oscar Newman's (1977) defensible space theory.The present study aims to evaluate existing literature and theories that relate criminality to climate to determine whether increasing urban expansion and densification,2 ,3 and diversity of people such as that which occur in ecotones (Kark 2013) is likely to lead to opportunities for an increased propensity for crime.Ecotones are where ecological communities overlap in urban areas, promoting diversity and beneficial exchanges.Factors like climate and community structure cause rapid environmental shifts in these overlapping regions (Holland et al., 1991;Kent et al., 1997).
The study aims to guide development policy and enhance urban design, policing infrastructure, and operational planning through a regenerative system thinking approach.It analyses crime types, climate factors, tree canopy cover, and population density, drawing from a systematic literature review and meta-synthesis of select studies.Gaps and limitations in previous research are addressed, emphasising the study's contribution and face validity, particularly in relation to heat and urban crime studies in highly urbanised areas, including Australian environments. 4his study contributes to the growing body of work on the climate-weather-crime relationship.This study differs from many other studies in the literature in several ways.It examines the relationship between criminality and climate, with a specific focus on the potential impact of loss of tree canopy due to urban expansion and densification on crime rates while also highlighting variables not yet considered in existing literature.Moreover, the study provides a unique perspective on this topic by focusing on a specific case study location, the City of Midland in Western Australia, and analysing the interplay between population growth, crime rates, demography, density, vegetation cover, and temperature.This case study approach offers valuable insights into the complex relationship between urbanisation and crime that are context-specific and could inform more effective urban design decisions and policing strategies.
Finally, the study uses a systematic literature review and meta-synthesis approach, which helps identify gaps and limitations in the existing literature and offers a more comprehensive and integrated view of the current state of knowledge in this field (Pai et al., 2004).Overall, the study's unique focus, methodology, and contribution to the literature make it a valuable addition to the existing body of research on crime prevention and urban design.

Urban heat island studies
Sustainability is a necessity for survival, therefore design strategies that improve sustainability outcomes take precedence over other political and economic motivations.Rad and Afzali (2021) highlight that urban morphology plays an important role in mediating immediate micro-climate and localised thermal comfort.Buildings are an example of this, and their form and orientation can further impact on urban heat island effect.Moreover, materiality, such as extensive masonry and road infrastructure is likely to exacerbate the condition as they add to thermal mass.Applying ENVI-met 3D simulations, Rad and Afzali highlights how different models of urban development provide improved conditions for thermal comfort over other alternatives.They found high-rise buildings affect micro-climate that impact outdoor thermal comfort.Chakraborty et al. (2019) found green vegetation in urban neighbourhoods moderate HIE.72% of cities sampled in their study showed disproportionate impact between affluent and vulnerable communities.An inference from this is that climatic events and the HIEs attributed to "urban structures which is then later radiated back into the environment" (Aguiar, 2012, 2), are not equitably distributed amongst communities.Aguiar's description suggests heat island conditions are aggravated in the evening due to thermal lag, or the rate at which heat that was absorbed by a material is released.High thermal mass such as hard surfaces that have absorbed solar radiation will retain the heat for longer and thus have a large thermal lag with implications for social, economic, and health related outcomes.Other studies cited by the author suggest a greener urban environment is gaining popularity amongst prospective urban dwellers.
Dew point is the interface between dry heat and humidity.This is important in the body's ability to cool itself because it is an element of thermal discomfort.Wet-bulb temperature (WBT) is the measure of heat and humidity in keeping with how a human might experience heat.5This is also important as it helps to evaluate the relationship between heat and humidity beyond human tolerance, as they interact in some parts of the world much faster than expected under global climate change (Raymond et al., 2020).O'Connor (2020) maintains "most climate scientists believed WBTs never rose above 35 °C, and never would until the effects of global warming grew stronger later in the twenty-first century".Sherwood and Huber (2010) extrapolate this further.They suggest a human body will overheat if heat and humidity combine to stop sweat from cooling the skin below 35 °C irrespective of health, level of activity, or whether in shade or not, and with unlimited drinking water supply.Therefore, WBT at this rate and lower can be uncomfortable and dangerous to health.
Heat, urban heat islands, crime behaviour, and the built environment Wolf et al. (2020) conducted a systematic review of 201 studies on the health impacts of urban trees, covering human health, environmental health, and urban forestry.The study highlights the potential of trees to mitigate urban heat island effects and reduce crime rates, particularly in low-income urban areas, especially for young individuals.However, the study acknowledges the presence of multiple variables that should be considered.The authors recommend integrating urban trees into urban planning and management to create health-promoting environments.It is important to note that the study is not specific to the Australian context.Nonetheless, applying the incivility and broken windows theories by Branas et al. (2011) in urban streetscape maintenance can offer advantages by reducing incivility, addressing visible signs of disorder, preventing crime escalation, and enhancing public safety.Municipalities can benefit from implementing these theories in their urban planning strategies.Wu et al. (2019) investigated the relationship between spatial access to urban parks and residential satisfaction in metropolitan Beijing.Their regression analysis indicated that residential satisfaction decreased with higher crime rates, and the interaction between park access and crime rates also had a negative impact on satisfaction.The study also found a significant association between distance to parks and perceived satisfaction, particularly in the "greenness domain," suggesting that closer proximity to parks led to higher satisfaction.The analysis considered variables related to residential satisfaction, spatial access to urban green spaces, neighbourhood characteristics, and household demographics, including age, gender, education, income, and family size.However, it is important to note that the study sample was more homogenous than the Australian context.While not directly aligned with the parameters of the current research, this study highlights the presence of other important environmental factors that influence behaviour beyond the scope of other studies.Baryshnikova et al. (2021) assessed weather-crime relationship at an hourly level rather than daily.At an hourly rate, precipitation was found to be a more significant variable on crime than temperature, and thus should not be disregarded.The research also studied the relationship between four other weather variables (temperature, humidity, precipitation, wind speed) and crime.Crimes included in the study are a select group of violent crimes and property crimes and city-level data, collected over 2014-2017 from Chicago, Indianapolis, Los Angeles, and New York, the largest cities in the USA by population.Findings suggest that the relationship between weather and crime differ significantly and is non-linear when observed at daily versus hourly intervals, and their subsequent impact on results and conclusion are relevant.The study, though not in the Australian context, further supports the notion that other variables are significant and must be considered.

Climate and crime
The literature suggests that the decrease in availability of green spaces leads to increased local temperatures and crime (Baron & Bell, 1976).This study supports the negative affect escape theory proposed by Baron (1972), which suggests a positive correlation between temperature and aggression until discomfort is reached at temperatures exceeding 30 °C.However, the generalisability of studies across hemispheres and socio-demographic groups is inconsistent.For example, recent research conducted in South Africa, a region known for its high crime rates, found that an increase in green space was associated with reduced rates of violent and property crimes, even after controlling for socio-demographic factors (Venter et al., 2022).This highlights the need for more comprehensive studies on the relationship between green space and crime, especially in urban areas of the Global South.
In Australia, the occurrence of major holidays during the summer draws attention to increased alcohol consumption and a subsequent rise in assault rates.It is important to note that the relationship between temperature and assault rates varies spatially and should not be generalised at a state level.This study also considers age and gender of both offenders and victims as variables, highlighting the potential impact of high temperatures on proactive policing due to reduced interactions and arrests.
Some studies highlight a positive correlation between the age of dwellings and suburbs with tree canopy coverage and socioeconomic markers (Donovan & Prestemon, 2010;Stevens et al., 2024;Troy et al., 2012).They also link this to population vulnerability, and propensity for crime.However, other disparate and stratified studies focus on specific locations and homogenous groups.To control variables in these studies, many observed the variables impacting on urban landscape, density, and crime relationships (Nieuwenhuijsen, 2020).Extant literature that examines northern hemisphere case studies suggest vegetation influences crime in both positive and negative ways.Examples of positive impact include mitigating the likelihood of crime through the cooling of the immediate environment and community temperament, the appearance of maintenance, and the perception of care.Examples of negative impact include increasing the likelihood of crime when it provides subterfuge.These are discussed further in this section.Muggah (2021, 3) reports a non-linear association between variables of climate and crime.Their observations include climate-related stressors that cause significant variances in violent and non-violent forms of crimes "ranging from homicides and assaults to robberies and burglaries".Such weather events have become more commonplace, and although the extent to which these links apply varies, Anderson and Anderson (1984), Anderson et al. (1997) and Rotton and Cohn (2003), who linked climate-related stressors and extreme temperature events with crime, and distinguished between "short-term shocks (cyclones, tornadoes, forest fires and floods), and long-term stressors (droughts and sea level rise)" (Muggah, 2021, 3), generally imply that rising temperatures can cause a significant rise in crime.Ranson (2012) attests that the profound impact of rising temperatures on criminal behaviour paints a grim picture based on a survey encompassing over 3,000 counties in the United States, forecasting a significant surge in criminal activity by 2100.The study predicts an alarming increase, including an estimated 35,000 additional homicides, 216,000 more rapes, 1.6 million further aggravated assaults, 409,000 additional assaults with a deadly weapon, and 3.1 million annual robberies.Citing the Intergovernmental Panel on Climate Change (IPCC) climate scenario, Ranson argues that the United States is poised to witness a notable rise, encompassing various offences such as murders, rapes, aggravated assaults, robberies, burglaries, larcenies, and vehicle thefts, compared to the baseline period of 2010.Further, Harp and Karnauskas (2020) evaluated 42 global climate models across 16,000 US cities and estimated an increase of between 2.3 and 3.2 million more violent crimes by 2100.These events can be attributed to high-density populations and locations susceptible to long-term climate stressors.
Many vulnerable communities are already impacted by extreme weather events and are dealing with the resultant severe economic and social impacts.Inequities and the ability to adapt exist across geographic locations.Examples include District 4 Vietnam, Copenhagen, Calcutta, Venice, Vancouver, The Maldives, The Torres Strait, and Australian states such as New South Wales and Queensland (Moosa et al., 2020).These are but a few amongst many more which underscore community divisions relating to approach, response, and adaptation issues, highlighting the challenge of seeking unification in mitigation planning (Jonescu et al., 2024).Notwithstanding this, scalable studies at the individual, local, and regional levels remain critical in Australia, even as most populated metropolitan centres are in proximity to areas most vulnerable to imminent pressures of extreme weather events.To this end, further cross-sectoral leadership will be necessary to coordinate an industry-led collaborative research approach and action a cohesive adoption of adaption strategies to enable a resilient and sustainable approach at local and national levels.Mares and Moffett (2015) assessed the link between climate change on levels of violence across 57 countries, sampling both western and non-western countries with regression results suggesting an increase in 1 degree Celsius in annual temperatures equated to an approximate 6% average increase in homicides.The authors later examined the impact of climate change on crime at specific times of the year.They utilised monthly crime statistics in the United States set against historical weather data that were collected by the Global Historical Climatology Network.Findings from these highlight a positive correlation between climate and crimes, with evidence supporting monthly variance clearly (Mares & Moffett, 2019).Stevens et al. (2019) conducted a study in New South Wales, Australia, analysing an 11-year dataset to investigate the influence of season and temperature on crimes such as assault, theft, and fraud.Their linear regression model indicates a positive association between daily assault and rising temperature.The study compares assault patterns with social media anger in response to temperature variations.While the study has limitations, including the comparison of different populations, it concludes that the routine activity theory applies (Cohn & Rotton, 2003), with aggressive crimes increasing in warmer weather, on weekends, and during holidays.The study suggests that increased outdoor socialising may limit opportunities for online presence and vitriol.Further research is needed to explore the relationship between temperature, specific activities, and resulting behaviour (Stevens et al., 2019).Stevens et al. (2024) investigated the impact of UHIs and green space on violent crime rates in Greater Sydney, Australia, between July 2013 and June 2018.They found that higher surface UHI was associated with increased rates of violent crime committed outside, while greater percentages of vegetation were linked to decreased rates of outdoor violent crime.These findings underscore the significance of addressing both violence and heat exposure, suggesting that expanding green spaces and reducing UHI could help mitigate these challenges exacerbated by urbanisation and climate change.
Amati et al. ( 2017) undertook a study across 139 metropolitan local government areas in Australia between 2015 and 2016.They established an estimated canopy cover and several socio-economic disadvantage benchmarks.Their calculation highlighted "heat island intensity" using satellite imagery, where they identify "hotspots" or heat islands across Australia through readily available data.The study also developed a vulnerability index that shows areas of socioeconomic and health disadvantage which coincide with a lack of green cover and a high incidence of heat.Notwithstanding this, both global and local analyses have demonstrated that median household income is one of the strongest determinants of urban crime incidence with other demographic variables such as homogeneity, prevalence of specific typologies (such as tobacco retailers), and the number of abandoned properties also positively associated with urban crime (Yu & Fang, 2022).Kuo and Sullivan (2001) also undertook studies on Ida B. Wells, a large highly densified public housing development in Chicago.The case study shows a "homogeneity" representation that is disproportionately high, with a lack of diversity evidenced in female residents (65%), African American (97%) and unemployment (93%).In addition, the study also sought to evidence a link between vegetation and crime, whether the former persuaded criminal behaviour based on passive surveillance and sentiment of ownership and safety.This is an important addition to knowledge as other research has only sought to link vegetation to urban heat islands where the heat itself is the subject of study in its effect on criminal behaviour.Donovan and Prestemon (2010) also consider the impact of vegetation density by relating it to limiting visibility and natural surveillance, in ways that advance the conversation on perceived safety through prevention e.g., inclusion of street lighting, security bars and systems, and the physical structures themselves.
The environment, psychological health, and safety are examined in Branas et al. (2011).The authors used incivility and broken windows theories to frame their research.They discovered that health and safety results relate to perceived attitudes of managed urban greening of empty spaces in Philadelphia, cared for and considered as "orderly" (as though someone had "power" over them).Greening's physiological (biophilic) and temperature effects on health and safety were not examined in the study.However, some associations imply major statistical decreases in stress and some crimes, while others show the contrary, such as disorderly conduct, which grew dramatically, and unlawful dumping, which only increased in specific city regions.
Poorly constructed urban places where people are exposed to a full spectrum of human behaviours constitute a safety risk.People avoid poorly lighted, low-traffic places that may attract crime (Jacobs, 1961, 89).Oc and Tiesdell (2007) argue that public spaces become socially sustainable due to increased occupancy and activity, passive surveillance, and extended habitation through "fortification approaches".These involve physical segregation and territorial defence, spatial governance within fortified and controlled spaces, and "regulatory approaches" that regulate public spaces and protect local municipal assets.Jonescu et al. (2018) demonstrate how surveillance technology might be integrated into urban environments to govern public places.These important planning principles reduce crime by encouraging occupancy, territorialisation, and social involvement in urban places.They also deter inhabitants from disruptive behaviours, vandalism, and other minor offences that affect personal and spatial safety.However, permeability, sightlines, and visual access are necessary for beneficial community interactions and social behaviours that improve safety perception and reality (Jacobs, 1961;Newman, 1977).Newman (1977) argues that a sense of ownership over a person's immediate urban environment empowers them for greater accountability and responsibility.In addition, according to Oc and Tiesdell (2007), in an "animated approach" to urban design, public spaces must be socially sustainable through increased use, passive surveillance, occupation, and activation.When communal places lack critical components of good urban design, lighting and activities that encourage participants to venture outside and stimulate community interaction, safety perception becomes crucial.Passive monitoring of a section of the population will lead to an increased sentiment of safety for communities to act in coordination with collective power to prevent undesirable behaviours (Cozens & Grieve, 2009;Jacobs, 1961;Jonescu, 2013).
These elements address observations against key criteria outlined by Gehl (2011).The elements are protection, comfort, environment, urban density, level of activity, increased physical participation, decreased traffic noise, pedestrianisation, and occupation of spaces.Other authors also agree these variables shape fundamental principles for designing urban environments that are socially sustainable (Cozens, 2011;Jonescu et al., 2020).

Case study development
Overall, the case study location was determined by evaluating several key criteria that would provide appropriate data to meet the research question.Criteria included temperature, vegetation cover, population data, crime statistics, and factors such as population, demographic diversity, economic indicators, and the feasibility of conducting the study in the selected location.This study utilised the Australian Real Estate Search, Suburb Reviews, to analyse residents' sentiments in a cross-section of suburbs across Perth (Homely Group Pty Ltd 2023).The consideration of mixed feedback in our study is essential as it provides valuable insights into the multidimensional nature of urban liveability and community well-being.
Residents' sentiments, as highlighted by Branas et al. (2011), serve as a critical indicator of perceived safety and overall satisfaction with the built environment which can significantly impact residents' perceptions of safety and contribute to increased criminal activity.Sentiment serves as a valuable resource for informing targeted interventions and urban development initiatives aimed at fostering resilient and thriving urban environments that encourage urban social interactions that can lead to improved passive surveillance outcomes and perceived safety.
The ideal candidate was required to have a functioning 'township' with inclusion criteria comprised wellestablished suburbs with a mix of land use, including residential, commercial, industrial, urban spaces, public transport, civic infrastructure, and educational and healthcare facilities.Localities that had a police precinct, a town centre, and other critical infrastructure, such as hospitals and industries were highly regarded.In the process of township development, all suburbs were cleared.The study recognised the significance of reintroducing vegetative growth alongside the criteria outlined in Table 1.This regeneration primarily occurred over time.
Based on the reviews of various suburbs, the study chose Midland over others considered.Although Joondalup, Armadale, and Mandurah met the Township Assessment Criteria, Joondalup's recent establishment limited its vegetation cover maturity.Mandurah's coastal location and cooling effects due to its proximity to the coast made it an outlier compared to other locations.Armadale closely resembled Midland across most criteria but lacked sufficient population density.
Midland received mixed online feedback.Positive aspects mentioned included its scenic location near Swan Valley, historic buildings, proximity to Perth Hills, schools, hospitals, shopping centres, and an efficient public transportation system.However, concerns were raised regarding criminal activity and abandoned places resonate with the notion of environmental incivility, suggesting potential areas of improvement to enhance community safety and cohesion (Churchill et al., 2023;Stevens et al., 2024) and broken windows theory (Branas et al., 2011).Some reviewers expressed reservations about settling down or raising families in Midland due to drug usage and safety concerns during night-time walks.
Midland, established in 1832, experienced rapid expansion following the commencement of operations by the Midland Railway Company in 1886.Various developments were introduced, including the iconic Town Hall in 1923.Midland's revitalisation was influenced by the Midland Redevelopment Authority, which became part of the Metropolitan Redevelopment Authority in 2012 (City of Swan, 2022).At the time of the study Midland had a population of 6,335 and a median age of 39 and is spread across a land area of 4.2 km2 resulting in a population density of 1,508 persons per km2 (ABS, 2022).The Midland CBD has a population of approximately 1,142 people and covers an area of 2.05 km2.The local municipality (City of Swan) projects Midland will experience an 84.6 percent increase in population by 2051 (id Consulting Pty Ltd, 2021).The associated dwellings required is anticipated to increase in step with population growth; density may impact local crime, and increased hardscapes may result at the expense of canopy cover.
Moreover, the feedback from residents also sheds light on the complex interplay between environmental factors, social dynamics, and community outcomes.As demonstrated by Stevens et al. (2020) and Amati et al. (2017), perceptions of safety and the quality of the built environment are closely intertwined, with visible signs of neglect and disorder often leading to decreased community engagement and social cohesion, leading to more time spent outdoors in the urban context leading to improved (Cozens, 2011).Nonetheless, Midland was recommended as an investment hub, particularly for apartments, due to its proximity to amenities and ongoing redevelopment projects (Homely Group Pty Ltd 2023).
By acknowledging and addressing the concerns raised by residents, urban planners and policymakers can work towards creating safer, more inclusive communities that promote positive social interactions and enhance overall quality of life.
The study also reviewed temperature, vegetation cover, population, and crime statistics data for the case study location.

Method design
This study applies three research methods.Figure 1 presents a relationship diagram between the research methods applied in this study.The methodology aligns with current literature, employing a systematic literature review to contextualise the study within existing scholarship (Kitchenham, 2004).Archival analysis is conducted to understand longitudinal context, demographic, and historical data.While the literature acknowledges potential for both linear and non-linear relationships, the data in this study is not normally distributed.Therefore, statistical analysis focuses on identifying non-linear correlations between climate, temperature, vegetation cover, population, and crime statistics.
A review of extant literature reported in this study has provided an understanding of some of the variables to be analysed in the study.City of Midland is the case study location.In addition, the systematic review has presented correlations between published theory, climate, temperature, vegetation cover, population, and crime statistics.An objective of this study is to undertake statistical correlation in relation to patterns of criminal behaviours and deviations from thermal comfort.Such an intersection between heat and urban crime is important in highly urbanised areas generally, not least the Australian environment. 6Archival data were explored using cross-referenced weather station temperatures near the case study area with ACORN-SAT used by the Australian Bureau of Meteorology to track long-term temperature changes nationwide.Scope of capture includes Perth Metro, Perth Airport, and Upper Swan, all 4-15 km from Midland CBD.,7 ,8 CSIRO's Urban Monitor was utilised to measure vegetation coverage, using Mesh Block Vegetation and Tree Canopy Cover.Relying on Caccetta et al. (2015Caccetta et al. ( , 2018)), this technique uses "grass", "shrubs", and "trees" to characterise vegetation by height: < 0.5 m, 0.5-3 m, and > 3 m.Canopy coverage is provided in square metres.Further contextual studies are conducted of the Midland suburb including population, age distribution, population density and anticipated growth.The overarching purpose of this research is to test a hypothesis on the relationship between temperature and crime behaviours: H1: higher temperatures increase crime rates.Midland Police District (Fig. 2) is the defined catch-   Contextually, Family means partner, ex-partner, parents, guardians of children, and children who reside or regularly stay with involved parties.Non-Family crimes are acts against people that are not included in the Family bracket.In addition, there are sub-category classifications of the WA Police that this current research did consider in its analysis.There were five such sub-categories under OaPty_Sel.They include Burglary (Bug), Stealing Motor Vehicle (S_MV), Stealing (St)-including from motor vehicles, Property Damage (ProDmg) and Arson (Ars).WA Police's crime categories and sub-categories were analysed by total crime rates per 100,000 people.9Total population was based on the annual Estimated Resident Population (ERP) published by the Australian Bureau of Statistics (ABS 2022).The research assumes that population growth was linear between two consecutive yearsas shown in Eq. 1.
As shown in Fig. 3, the mean monthly temperature histogram had two peaks, due to the seasonal effects.Since this data distribution was non-normal, non-parametric statistical tools were used to test H1.

Monthly crime rates
Table 2 shows descriptive statistics for monthly crime rates for crime groups and sub-categories.Of the main groups, OaPty (µ = 529.22;σ = 71.25)has the highest average crime rates, followed by OaP (µ = 128.16;σ = 26.12).

Monthly mean temperature versus monthly crime rates
To test H1, this research set the null and alternative hypotheses as below: H 0 : there is no significant correlation between monthly crime rates and mean monthly temperatures.
H a : there is a significant correlation between monthly crime rates and mean monthly temperatures.
Table 3 illustrates only the statistically significant associations at the 95% confidence level (p ≤ 0.05).Three categories of crimes indicated significant statistical correlations.OaP shows the strongest correlation (r = 0.336; p < 0.001).The R-squared value shows about 12.5% of data fit the trendline.Drugs and RnP_SP indicate weak negative correlations with mean monthly temperature.Therefore, the association between crime categories and monthly temperatures for Midland District is limited.Following these observations, sub-groups of OaP and OaPty were analysed further.Table 3 indicates A_ nFam (r = 0.449; p < 0.001) and Ars (r = 0.485; p < 0.001) showed moderate positive relationships with monthly temperature.Since data distributions were non-normal, Table 2 Descriptive statistics for monthly crime rates for crime groups and sub-categories    Figure 4 shows about 20% of data fit the trendline.The positive correlation between Ars and temperature can be understood as hotter days leading to greater fire risks.This finding underlines the need to further study why higher temperatures increase assault rates towards nonfamily members.

Analysis of midland: vegetation cover, population and crime statistics
A report by the Department of Planning, Lands and Heritage (2021) shows Mesh Block Vegetation Cover and Tree Canopy for Midland in 2009 and 2016 increased from 7.98 to 10.39% of landmass.In addition, according to datasets of the ABS (2022), Midland's population of 6,335 (median = 39) is spread over a landmass of 4.2km 2 (density = 1,508 persons/km 2 ).Midland CBD has a population of approximately 1,142 people over a landmass of 2.05 km 2 .The population of the City of Swan municipality, in which Midland exists, is projected to rise by 85% by 2051.The number of dwellings required is predicted to increase sharply over the same period to keep in step with population growth.
Statistics show ± 20% or higher between the summer months of December to February, in comparison with the preceding 9-month average.

Discussion
This study emphasises the complexity and multitude of variables involved in understanding the relationship between heat, heat-related illness, and crime or violent behaviour.Studies by Stevens et al. () specifically investigates how seasons and temperature impact crime rates, focusing on variables such as assault, theft, and fraud.2023) investigated the impact of temperature and climate change on crime, uncovering possible long-term patterns and consequences for policy development and intervention tactics.Corcoran and Zahnow (2021;2022) examined the correlation between weather conditions and assault rates in Brisbane, and undertook a systematic review of the empirical literature, offering significant insights into specific patterns of criminal activity.In addition, Auliciems and DiBartolo (1995) conducted important research on the relationship between weather patterns and domestic violence in the same urban setting, emphasising the complex impact of environmental factors on criminal behaviour.
In contrast, other studies have primarily explored the relationship between heat and social media behaviour or health outcomes.Developing on these influential studies with our own research, we enhance our comprehension of the elements that influence crime patterns in urban areas of Australia.This highlights the need for more interdisciplinary research and collaborative efforts to tackle the complex challenges presented by climate variability and its consequences for community safety.
This research suggests that studies conducted beyond the United States and European contexts are relatively limited, making generalisation to other socio-economic, cultural, and ethnic circumstances challenging.Moreover, had significant limitations and fails to consider criminality comprehensively within the broader context of societal dynamics, social norms, and landscape morphology over time.
In Perth, tree canopy area increased by 4% between 2009 and 2016.Tree canopy coverage and suburban maturity have been widely researched (Lowry et al., 2012;Luck et al., 2009;Mockrin et al., 2019).Troy et al. (2007) found that residence age and tree coverage are positively correlated (around 45 years).Mockrin et al. (2019) found different relationships between dwelling age and urban tree canopy.These relationships were non-linear and negative when comparing United State locations.Lowry et al. (2012) added that community and housing age influenced the relationship between urban tree cover and socioeconomic factors of vegetation cover.Luck et al. (2009) found that housing density increases vegetation over time, showing that communities 'green up' after development.
Global studies also neglect "blue" infrastructure like coastal areas and vast bodies of water.Thermal mass and direction are also a vital omission.Most capital cities in Australia are located around the coast, which is colder and wetter than interior places.As air travels from highto low-pressure zones, cooler ocean air pushes toward warmer air onshore, generating a sea breeze.This cools coastal areas but not interior areas.Blue infrastructure was rarely included in heat-crime research, nonetheless a key variable in HIE.Urban development's influence on cooling channels may have substantial implications for understanding sea breezes in cities.
Research typically correlates environment, vegetation, social, and demographic information.Law enforcement efforts and initiatives are usually overlooked e.g., police "actions" that target certain forms of crime in specific locations.Guerette and Bowers (2009) found this may lower crime in targeted regions while increasing crime in non-targeted areas.Normative literature, e.g., Clarke and Weisburd (1994) has reported on crime implications of displacement and diffusion.Moreover, transit nodes are, and their surrounding environments that the extant literature suggest, are interconnected, and are largely not considered as variables in the heat-crime literature as it is in the urban-crime literature.
Moreover, studies do not typically attribute drug use and seasonal drug use patterns (alcohol) as a variable in the populations studied.For example, the impact of heat on drug-affected persons and crime.Surveys relating to illicit drug use by people aged 14 and over in Australia considered age, geographic area, level of socioeconomic advantage and disadvantage, and a person's education (Australian Government [Australian Institute of Health and Welfare] 2020).In 2019, 43% of people in Australia were reported to have used an illicit drug at some point in their lifetime, whilst 16.4% had used one in the last 12 months.For example, the use of cannabis, cocaine, ecstasy, inhalants, hallucinogens, and ketamine all increased from 2016 to 2019, and so was the non-medical use of pharmaceuticals, painkillers, opioids, and codeine.Future studies may draw conclusions on such increases and their relevance to variables of crime, urban tree canopy, heat and HIE.While our study did not specifically analyse seasons, it is important to recognise their potential impact on crime dynamics, especially in Australia.For instance, the summer season coincides with holidays and increased alcohol consumption, potentially influencing aggressive behaviour.Although a limitation in this study, acknowledging seasonal variations could offers valuable insights and highlights the need for future research to explore this aspect further, particularly in regions like Australia where seasonal patterns differ.
Crime statistics need to be considered proportionately and examined per 100,000.This is to examine the correlation between crime rate and population growth, dwelling growth, and growth in economic activities.In addition, offenders' statistics may include alleged offences that are later retracted or unsubstantiated, which are not redacted from databases and risk biasing results.In Australia, lesser incidents such as disturbances, disorderly behaviour, damage, and other events may be reported to municipal rangers instead of police.Such occurrences are not included in police crime statistics, and offences of a sexual nature, and domestic violence are more likely to affect women, children, and disadvantaged socioeconomic groups, regardless of vegetation density.Such crimes may be underreported due to apathy toward authority and sensitivity of the crime type.
Service and coverage may vary per municipality.The percentage of house ownership and rental might also impact findings; however, they are not uniformly examined among studies.Crime may also change with population density.Radial, suburban, and gridded patterns that create mono-functional zones or segregated precincts may catalyse crimes.High-rise residential with mixeduse, single residences, pedestrianised precincts and public realm, vehicular-dependent cities, rate of movement, and scale also affect sustainable densification and placemaking (Jonescu et al., 2018).
Prima facie, the Midland case study indicated a positive correlation between temperature, crime statistics, and green canopy cover (which rose over time) with population density.Thus, longitudinal studies might draw different conclusions across the literature reviewed.Accordingly, future studies could consider recidivism rates [a small fraction of the population overrepresented in crime statistics], in conjunction with population density.Heat, HIE, and canopy cover may not affect recidivism rates, which may have grown independently of population expansion and density.
While this study primarily focused on the relationship between temperature and crime rates, future research endeavours could explore the intricate interplay between rainfall patterns, temperature fluctuations, and their combined influence on criminal behaviour, thus providing a more comprehensive understanding of environmental factors impacting crime rates.This study acknowledges the importance of incorporating contemporaneous vegetation data in alignment with ongoing housing developments that may displace remnant vegetation in future studies, with relevance to presenting a deeper understanding of the multifaceted dynamics of urban development and their effects on vegetation loss and criminal behaviour.Additionally, exploring the potential influence of blue grids, such as water bodies, on temperature reduction and subsequent impacts on crime rates is an area of worthy of future research in this emerging area of study.Lastly, while this study did not incorporate emergent sources of disaggregate data, such as smartphone tracking studies, to analyse UHI exposures through the lens of everyday activity spaces, future research could explore the potential of integrating such data to enhance the scholarly understanding of this relationship.

Conclusion
Existing studies linking temperature and crime rates lack longitudinal analysis.This study highlights anomalies that warrant further investigation, including recidivism, urban composition, and variables beyond temperature, population, and crime.Factors such as economic downturns, unemployment, and apathy in reporting can influence crime rates.It is important to consider the variability in location-based climate change outcomes and the need for more comprehensive studies on the relationship between green space and crime, especially in urban areas of the Global South.
The study emphasises the heat-sink properties of thermal mass which can lead to exacerbated UHIEs, orientation, which can influence passive cooling, and blue grids which, in conjunction with orientation and breezes can be utilised to alleviate heat, are often ignored in normative literature, in the design and development characteristics of cities researched.And further, the potential significance of integrating rainfall analysis into future research endeavours, which could offer valuable insights into the multifaceted influences shaping criminal behaviour.Urban morphology mediates immediate micro-climate heat islands with dramatically different temperatures from surrounding local belts.
Human behaviour and criminality are complex, posing challenges for urban design decisions.The impact is further complicated by changes in macro-and microclimates.Urbanisation and the loss of green spaces, leading to urban heat islands, are also significant factors.The existing literature is still in its early stages, mainly focused on Northern Hemisphere case studies.However, if hypotheses regarding the interplay between the urban environment, local climate, and criminal behaviour are accurate, it is crucial for built environment specialists to collaborate on evidence-based research to understand the intricacies and determinants of emerging crime and behavioural norms.This collaboration aims to mitigate the consequences of urbanisation.This study's analysis of growth, crime, demographic, and temperature analysis of Midland, in Perth Western Australia assessed the validity of existing theories that associate criminality and patterns of criminal behaviour.Local crime, temperature, and vegetation cover data were analysed.Validity was done through an evaluation of field literature.
The findings draw to an observable correlation between green canopy cover with temperature, long-term crime, and population density.However, the population for Midland was expected to grow substantially by 2051.The study acknowledges that while the commonly tested variables applied by most other studies did not have a statistically significant impact on crime in the case study area, there are other variables that may influence crime discussed throughout the paper.
Such prediction of the potential impact of local heat and canopy cover on urban criminal behaviour could benefit the long-term aspirations of designers and policymakers in the (re)shaping of city infrastructure through development policy largely set prior to the eminence of climate change and heat islands.Development policy, for example, could consider alignment, orientation, above and below-ground infrastructure, materiality, blue grid optimisation, vegetation, and greening policies for active cooling.Moving forward, the integration of rainfall analysis as a complementary variable holds promise for refining predictive models and informing policy interventions aimed at mitigating crime rates in urban environments, and further research in police resource planning and resource implementation could benefit from understanding the likely correlations between long-term heat projections and an anticipated increase in crime.
Additionally, this study highlights the importance of considering environmental variables such as vegetation cover and temperature in understanding crime dynamics.We recognise the need for future studies to include more up-to-date vegetation data available, especially in areas undergoing housing development, and suggest prioritizing this in future research endeavours.
Lastly, the potential impact of blue grids, such as water bodies, on temperature reduction and crime rates warrants further investigation.Incorporating such variables into future studies can enhance our understanding of the complex interactions shaping urban crime patterns, offering valuable insights for policymakers and urban planners.

Appendix
ment of the study, consisting of three local government areas: City of Swan, City of Kalamunda, and the Shire of Mundaring.Mean monthly temperature data for the study were extracted from the Bureau of Meteorology (2022), with Perth Airport Station Number 9021 being the closest weather station.Crime statistics from the WA Police archives were extracted from the Western Australian Police (2021).They cover records of offences that are known to Midland Police within the relevant periods, excluding incidents categorised as no criminal offence, offence substituted, false report, mistakenly reported, or entered in error.Monthly crime statistics were taken from January 2007-December 2020 (data for the study including the year 2020 was treated with caution because of COVID-19 pandemic and the social restrictions that came about in controlling it).All data obtained for the purpose of the study are open source and publicly available from: City of Midland, Australian Bureau of Meteorology, CSIRO's Urban Monitor, and Western Australian Police Force.Data classification Crime classification is by the Australian and New Zealand Standard Offence Classification (2011), WA Legislation, and operational reporting requirements (Western Australian Police, 2021).From WA Police

Fig. 1
Fig. 1 Diagram of research methods and their relationships

Fig. 4 a
Fig. 4 a Scatter plot of AoP against mean monthly temperature.b, c Scatter plots between A_nFam and Ars, and mean monthly temperature Between 2007 and 2020, Western Australian Police (2021) figures suggest that this occurred on three occasions for AoP (the average from March 2012 to November 2012 contrasted with December 2012 to February 2013 (+ 25%)).Also, between March 2014 and November 2014, set against December 2014 to February 2015 (+ 21%), and averages across March 2018 and November 2018 compared with December 2018 to February 2019 (+ 22%).This did not occur for AoPty but on one occasion, the average of March to November 2010 when compared with December 2010 to February 2011 (+ 19.6%).Across most other similar comparisons in this category, most increases are of less than 5%, and on three occasions had reduced between negative −0.5 to −10% (March 2009 to November 2009 compared with December 2009 to February 2010 (−10%)).It is also important to note that in Australia, December to February coincides with school holidays, which also coincides with increases in some crime categories.

Table 1
Township assessment criteria across six case study candidatesCriteria that was shared equally among all candidates was not included in the table 6Due to the unavailability of crime data at the desired frequency, it was not possible to compare daily air temperature data with the timing of the crimes.

Table 3
Spearman's correlation coefficients between mean temperature and crime rates