Changes in storm characteristics over time can significantly impact such hydrological processes as stream flow and soil erosion. During a project to update the precipitation climatology in the Revised Universal Soil Loss Equation (RUSLE), a number of storm characteristics were derived from the NCDC 15-minute precipitation data (TD-3260). Linear regression was used to identify trends in storm precipitation totals, storm duration, storm intensity, maximum 30-minute amounts, and total storm erosivity. Unlike most studies, this one will focus on actual storms instead of daily totals. A storm is defined as a period of precipitation separated from other precipitation periods by more than 6 hours. Results will be presented for annual and seasonal trends. Analysis suggests that the patterns in the seasonal trends of storm characteristics are more coherent than in the annual trends. Furthermore, the interaction among the variables can be complex. For example, Texas in spring shows a decrease in storm intensity (amount/time) that is not driven by reduced storm amounts but instead by increased storm duration.