Abstract
Ambient temperature has been demonstrated to be associated with a variety of violent or conflict events. However, few studies have so far linked temperature to the risk of police violence, where the quantitative estimate of the temperature effect is still unclear. In this study, we comprehensively explore the relationship between temperature variations and police violence risk based on a series of panel regression models with high-dimensional fixed effects in the United States. The results indicate a generally positive association between temperature and the police violence, with higher temperatures corresponding to elevated risks. The heterogeneity analysis exhibits that lower levels of precipitation and larger population sizes are associated with increased risks of police violence. Specifically, under the conditions of less than 50 mm precipitation and a population of larger than 5 million, each 1°C rise in monthly average temperature is associated with an increased death rate of 2.06 (95% CI: 0.92-3.20) and 2.01 (95% CI: 1.08-2.93), respectively. The temperature effect on police violence risk presents notable spatiotemporal variation, with elevated risks observed in certain states with experiencing high temperature and particularly during the year of 2024. Our research projects that by the year 2050, under the highest greenhouse gas emissions scenario of SSP5-8.5, the cumulative additional deaths from police violence in the United States due to expected temperature change would achieve 479 (95% CI: 183-836). Given the profound and widespread societal impact of deaths related to police violence, these projected additional deaths may pose further challenges to public health and social stability in the United States. Our research reveals the linkage between temperature variation and the risk of police violence, highlighting the urgent need for targeted intervention strategies in the practices of police law enforcement, particularly under the high-temperature environmental conditions.