A plethora of studies have found how certain types of weather patterns promote violence (see Rotton and Cohn 2003; McDowall et al. 2012; Mares 2013; Xiaofeng et al. 2017, Stevens et al. 2019). With few exceptions (see Ekwall and Lantz 2015; Lynch et al. 2022), the general consensus among scholars is that an increase in temperature in particular is associated with an increase in crime, all things considered (see Cohn and Rotton 1997; Larrick et al. 2011; Ranson 2014). Studies have most often found that the degree of association between temperature and crime fluctuates based on a range of factors including the type of crime, the temporal and spatial scale, as well as the geographic locale under investigation. For example, an increase in temperature has most often been found to be positively associated with violent crimes such as assault (Breetzke and Cohn 2012; Corcoran and Zahnow 2021) and homicide (McDowall and Curtis 2015), but not necessarily with property crimes (Cohn 1990; Ding and Zhai 2021; Lynch et al. 2022). The notion here – aligned with routine activities theory - is that an increase in temperatures increases opportunities for socialisation increasing the opportunities for various types of interpersonal crimes.
Temporally, studies have examined the crime-temperature linkage at a range of resolutions including annually (Rotton and Cohn 2003; Szkola et al. 2021), monthly (Mares 2013; Churchill et al. 2023), weekly (Ceccato and Uittenbogaard 2014; Jung et al. 2020), daily (Schinasi and Hamra 2017; Schutte and Breetzke 2018) and hourly (de Melo et al. 2018; Towers et al. 2018) with no discernible gradient in strength of positive association across scale. An increasing number of studies have begun to examine the temporally lagged effect of temperature on crime (see Xu et al. 2020; Ankel-Peters et al. 2023). For example, Tiihonen et al. (2017) found a one-month lag effect for high temperatures on violent crime in Finland while Potgieter et al. (2022) found a seven-day lag effect for high temperatures on the relative risk of violent crime in South Africa. The causal pathways linking weather to an increase in crime over time are myriad and are thought to be associated with how weather shocks can disrupt people’s daily routines, leading to increased stress and frustration and increased crime risk (in the short term); and to increase tangential crime risks such as increased inequality (Kelly 2000; Manea et al. 2021), and/or lower agriculture productivity (Somanathan et al. 2021), among others (in the long term).
In terms of spatial scale, the vast majority of research has found positive associations between temperature and violent crime regardless of the spatial scale under investigation. By far the most common spatial scale has been citywide (see Gamble and Hess 2012; Reeping and Hemenway 2020; Xu et al. 2020), followed by district (Berman et al. 2020; Wu et al. 2020) although an increasing number of studies have investigated this association at the neighbourhood-level and below (see Jung et al. 2020; Gorislavsky and Mares 2021). From a location perspective, there has been a tendency for the weather-crime research to focus on Western contexts that experience temperate climates (Corcoran and Zahnow 2022), although studies have now emerging from lesser studied regions of the world including Saudi Arabia (Algahtany et al. 2023), India (Blakeslee and Fishman 2018), Columbia (Trujillo and Howley 2021), and Nigeria (Afon and Badiora 2018), among numerous others. Regardless of the spatial scale or geographic locale, the overall picture that has emerged from this growing body of literature is that higher temperatures increase the probability of interpersonal violence, and criminal behaviour.
Methodologically, researchers have typically applied cross-sectional study designs (Rotton and Cohn 2003; Ranson 2014; Reeping and Hemenway 2020), and/or used longitudinal data and employed time-series designs to examine the impact of weather on crime (Schinasi and Hamra 2017; Hu et al. 2024). In the majority of these studies a mean (daily, weekly, monthly, annual) temperature is calculated and compared with those during control periods in the same location under study. More recently however, studies have examined the impact that anomalous and/or disproportionate variations in temperatures (or temperature ‘shocks’) have on crime rates (see Jacob et al. 2007; Gangopadhyay and Nilakantan 2018). The notion here is that while broader weather fluctuations may, over time, increase crime risk it is rather extreme and intense spikes in unseasonal weather that may have an acute impact of crime. The rise in the occurrence and destructive power of extreme weather events such as extreme heat or rain is one notable by-product of climate change that is likely to impact various forms of behaviour, including criminal behaviour. That is, extremely hot temperatures will likely increase the discomfort and frustration levels of individuals leading to more aggressive behaviour and ultimately more violent crime. This causal mechanism is substantiated theoretically by the heat-aggression mechanism (Anderson 1989), which motivates that hot temperatures increase aggressive motivation and (under certain exigent conditions), aggressive behaviour.
Defining what constitutes an ‘anomalous’ weather period is variable however with most researchers employing a ‘departure-from-historic-normal’ approach to delineate a weather extreme or volatility. For example, Mares et al. (2013) deducted the long-term mean temperature monthly value (from 1971–2000) from the actual mean monthly value and extracted anomaly values whereas Churchill et al. (2023) calculated temperature shocks as the difference between the observed temperature and the long-run mean divided by the long-run standard deviation. Other studies have extracted and/or combined the hottest days per year over a five-year study period and compared crime on these days with ‘normal’ or ‘random’ days (see Schutte et al. 2021) or compared daily temperatures departing from more recent temperatures to capture an element of unexpectedness and measure weather volatility (see Thomas and Wolff 2023). Regardless, in the vast majority of instances the positive association between temperature and crime holds true.
Studies examining the impact that rainfall specifically has on crime has generally been less forthcoming but studies that have been undertaken have most often produced mixed results with some finding an increase in crime during periods of increased rainfall (see Hsiang et al. 2013; Shen et al. 2020) and others finding no such association (see McLean 2007; Schutte and Breetzke 2018). The operationalisation of rainfall in studies is usually more straightforward and has been measured as the presence or absence of rainfall on any given day, week or month and compared rates of crime on comparable days in which no rainfall is recorded (see Mares 2013; Linning et al. 2017; Sommer et al. 2018).
Despite a lot being known about the weather-crime relationship there are still some notable gaps in the extant literature. That is, most previous studies examining this relationship have been undertaken across single cities (see Jung et al. 2020; Reeping and Hemenway 2020; Shen et al. 2020), or across a large number of cities within the same country (see Berman et al. 2020; Xu et al. 2020; Peng and Zhan 2022; Thomas and Wolff 2023) with much fewer studies examining this linkage across countries and none, to our knowledge, having examined this association across cities in different countries but within the same broad geographic and climatic zone. The small number of global empirical studies that have been undertaken have largely corroborated the weather-crime linkage, but the results are mixed and nuanced. For example, Rotton (1986) found positive correlations between homicide and a number of climatic variables across 41 countries but found that these positive associations disappeared when a social variable (life expectancy) was partialled out of the analysis. Similarly, Lynch et al. (2020) found that the long-term positive correlation between temperature and homicide in New York and London (for 1895–2015) disappeared when gross domestic product was controlled. Other studies have found weather to have an indirect effect on homicide, mediated by other factors. For example, Barlett et al. (2020) connected global warming to extreme weather events that threaten clean water supplies, which, they inferred, creates resource stresses that motivate criminal behaviour. Finally, Kuznar and Day (2021) examined the causal connections between homicide, inequality, and temperature in 173 countries and found considerable variability in the extent to which these three phenomena combine. Specifically, they found homicide rates to be higher when poorer segments of populations were disproportionately influenced by temperature. Other cross-national studies include Mares and Moffett (2016) who found that each degree Celsius increase in annual temperatures in a sample of 57 countries was associated with a nearly six percent average increase in homicides and Wei et al. (2022) who found a direct and positive relationship between higher temperatures and homicide for 171 countries from 2000 to 2018. The results of these collective cross-national work suggest that examining the weather-crime linkage across contexts is not straightforward but rather is complicated by mediating factors unique to each locale.
In this study we add to what is known in this scholarly space by examining the impact that anomalous temperature and rainfall days have on violent crime in two diverse cities on two separate continents but that are situated in similar climatic contexts. We hypothesise that anomalous weather will be associated with significant shifts in violent crime regardless of context. Findings from this study can potentially be used to develop and inform effective crime prevention strategies for policing in both contexts during extreme weather events and highlight the importance of climate change regulations in communities.
Study sites
Khayelitsha
The township of Khayelitsha is located approximately 30 kilometres to the south-east of the Cape Town central business district (CBD) in a region colloquially known as the ‘Cape Flats’. The Cape Flats has a unique history in the country. During apartheid, the national government forcibly relocated all non-White communities to the Cape Flats, leading to the creation of so-called ‘townships’ which were (and still are) characterised by high levels of poverty, overcrowding, and inadequate service delivery and infrastructure. Paradoxically, the Cape Flats became a centre of resistance against apartheid, witnessing significant political activism and struggle. Since democracy in 1994, all townships within the Cape Flats have continued to face socio-economic challenges, including rising poverty, high unemployment, and crime, particularly gang violence (see Pinnock 2016). Khayelitsha is among the poorest townships on the Cape Flats with roughly 40% of residents unemployed, and youth unemployment (aged 15–23) at over 50% (Statistics South Africa (SSA) 2011). The township has a population of roughly 400,000 inhabitants (SSA 2011). Crime is rampant in Khayelitsha with the main policing precinct in the township consistently among the most violent in the country with contact crime almost double the national average (Crime Hub, 2023). In terms of climate, Khayelitsha experiences a moderate climate with mild, wet winters and dry hot summers. The summer season (December to February) brings warmer temperatures, ranging from around 15°C to 28°C (59°F to 82°F), with little to no rainfall and occasional hot days. During the winter months (June to August), temperatures range from around 8°C to 18°C (46°F to 64°F), with occasional rainfall and cool evenings.
Ipswich
Ipswich is a city located approximately 40 kilometres west of the Brisbane CBD in Queensland, Australia. The city has a diverse population of roughly 250,000 inhabitants (Australian Bureau of Statistics 2022) the majority of whom have an Anglo-European background. Recent decades have, however, seen significant demographic shifts, with increasing cultural diversity due to immigration from countries such as Vietnam, India, and the Philippines. The population of Ipswich is relatively young, with a significant proportion under the age of 30, and the city is known for its affordable housing compared to neighbouring urban centers like Brisbane. According to the Queensland Police Service (QPS 2024), the crime rate in Ipswich is considerably higher than the average across most suburbs in Australia. In fact, Ipswich is ranked in the top two percent of the neighbourhoods across Australia in terms of crime risk. The city typically experiences a subtropical climate characterised by hot, humid summers and mild, dry winters. During the summer months (December to February), temperatures can often soar above 30°C (86°F), occasionally reaching into the mid-40s Celsius (over 110°F) during heatwaves.