A new technique was recently developed to study seasonal cycles of climatic trends of expected values, variance, skewness, and other statistical moments of climatic variables. Here we apply that technique to analyze the diurnal and seasonal cycles of trends of surface air temperature and its variability using hourly observations from nine geographically distributed meteorological stations in the United States for the period 1951--1999. The analysis reveals a complex pattern of trends in temperature and its variance at different times of the seasonal and diurnal cycle, showing warming trends for all stations during most times of the year and times of day, but with diurnal asymmetry of warming only in the warm half of the year. We found no correspondence between the trends in temperature and the trends in temperature variability. This analysis may be used as a prototype for developing the next generation of climate services that will be able to supply customers with detailed information about the first few moments of the statistical distribution of any meteorological variable for every day and hour of the period of observation.