. . . [P]leasant evening weather may increase the number of opportunities for mugging, and dark, rainy nights may increase the probability of successfully burglarizing a house without being detected.
A second explanation draws on a social interaction theory of crime. Glaeser, Sacerdote, and Scheinkman (1996) propose that the frequency of criminal acts is driven in large part by social interactions that occur during day-to-day life. Applied to weather, such a hypothesis implies that weather conditions that foster social interactions are likely to increase crime rates (Rotton and Cohn, 2003). For example, mild weather that encourages people to go shopping would also have the effect of increasing the frequency of property crimes such as larceny.A third possible explanation draws on theories in which external conditions directly affect human judgment in ways that cause heightened aggression and loss of control (Baumeister and Heatherton, 1996; Card and Dahl, 2011).Experimental evidence strongly suggests that ambient temperatures affect aggression (Anderson, 1989). For example, Baron and Bell (1976) assigned male subjects to receive a positive or negative evaluation from a confederate, and then gave them the opportunity to retaliate with an elecric shock. They found that retaliation was highest when the experiment took place in a room with a high ambient temperature (92-95 degrees F), and that retaliation was still heightened even at more moderate temperatures (82-85 degrees F). Such studies imply that weather may directly influence people’s psychological propensity to commit violent criminal acts.
The bottom line is that increases in temperature lead to increases in crime. See the paper by Matthew Ranson here (July 2012).