Harp and his advisor, CIRES Fellow Kris Karnauskas, used climate analysis techniques to investigate the relationship between year-to-year fluctuations in climate and violent crime rates in U.S. cities since 1979. Their methods accounted for the fact that crime rates have dropped significantly since the 1990s in most places. These long-term trends, driven by many societal factors, create the “baseline” for the new analysis. “Consequently, we considered the crime rate differences from that baseline,” said Karnauskas, who is also an Associate Professor of Atmospheric and Oceanic Sciences at the University of Colorado Boulder.

He and Harp obtained monthly violent and property crime data for more than 16,000 cities directly from the FBI, specifically the Uniform Crime Reporting (UCR) Program. The database included all types of violent crimes including murder, rape, robbery, aggravated assault, burglary, larceny, and motor vehicle theft. The scientists relied on historical climate data from NOAA's North American Regional Reanalysis (NARR).

They divvied data up by broad climate regions in the United States, and then measured the strength of the relationship between climate variables and crime in each region.

Combining these data sets revealed a strong relationship between crime and temperature, in particular, including a much stronger correlation in winter than in summer months. For example, in winter in the northeastern United States, the relationship was so tight that temperature changes alone could explain more than half of the year-to-year ups and downs in crime rates. In the summer, the relationship between temperature and crime patterns diminished.

The strength of the wintertime correlation was surprising, Harp said, given that crime rates vary for all kinds of reasons.

“It's highly unusual to find correlations this high in big, messy data sets, especially spanning disciplines like climate and health or sociology. The initial disbelief forced us to recheck our work more than a couple of times,” Karnauskas added.

The new assessment also provides insight into why climate anomalies affect crime rates, including some of the strongest evidence to date in support of one theory about how crime patterns may be linked with weather and climate: The Routine Activities Theory that states that despite the complexity of human behavior and external forces, interpersonal crime is driven by a relatively simple combination of ingredients: a motivated offender, a suitable target, and the absence of a guardian who could prevent a violation. So pleasant weather may increase the chances of all three factors converging; lousy weather may decrease it.

“This study is significant because it broadens our thinking on connections between climate and human health, to encompass a very real and dangerous threat to our bodily safety and, therefore, health,” said Karnauskas.