A decade ago, Harold Brooks and his collaborators discovered that the number of U.S. tornado outbreaks per year have been gradually increasing in time, yet the number of days on which tornadoes actually occur have been gradually decreasing in time. The key implication of these dichotomous trends is that more tornadoes have been forming on a fewer number of days. Motivated by this discovery, we asked two questions: When during the calendar year, and where geographically, are these trends the most pronounced? Using a database of the most recent 63 years of tornado data, we found that the daily tornado activity has decreased rather strikingly during the months of June through August, primarily within the Southern Great Plains region of the United States. In contrast, we found that the frequency of tornado outbreaks has increased significantly during warm- as well as cool-season months, particularly within the Southeast U.S. The temporal changes have resulted in a three-week shift in the peak probability of a U.S. tornado day, from the middle of June to the last quarter of May. The spatial changes are of consequence because of the documented high tornado vulnerability of communities in the Southeast U.S. Finally, the continued increase in tornado outbreak frequency over the most recent decade is of concern because outbreaks tend to include tornadoes of high-end intensity and have a tremendous impact on life and property. We recommend that future efforts to understand the range of tornado-outbreak predictability be prioritized.