Effect of temperature and precipitation change on crime in the metropolitan area in Virginia, USA

Abstract

ABSTRACT Previous climate-crime studies are primarily grounded on conventional statistical methods, leaving the possible non-linear nature of the climate-crime association insufficiently explored. In the present study, we employed both multiple linear regression and wavelet analyses and compared their results to detect and verify the possible non-linear effect of temperature and precipitation change on crimes in an urban setting. Nine types of crimes in Hampton Roads, Virginia, USA during 1973–2009 were analyzed. Results show that temperature was the significant determinant of crimes in both of the analyses. Wavelet coherency analysis demonstrates that temperature and crimes exhibited common modes of oscillation in ∼1-year periodicity, which may be attributable to the seasonality of crimes. We further found that the seasonality of crimes was determined by the seasonal change of temperature rather than that of precipitation. In parallel, precipitation and crimes exhibited common modes of oscillation in ∼11-year periodicity, revealing their association at the decadal scale. Our results indicate that temperature and precipitation determine crime rates at different time scales. Nonetheless, when the crests of the temperature-crime and the precipitation-crime cycles overlap, crime rates may be reinforced to their greatest extent. The above findings may help inform and prepare necessary security arrangements in advance. Further research could be conducted to see whether the above findings could be applied to the metropolitan regions with the similar geographic settings of Hampton Roads, such as the Jing-Jin-Ji (Beijing-Tianjin-Hebei) region in northern China.