We present a new technique to analyze diurnal and seasonal cycles of climatic trends in records with arbitrary observation times. This technique produces unprecedented opportunities to analyze and visualize climatic change. There is no restriction that the climatic process be stationary, and records can contain seasonal and diurnal cycles in the moments of the statistical distribution (means, variances, and higher moments) of the observed variables and their trends. To approximate such processes we use superposition of a limited number of Fourier harmonics of diurnal and annual cycles with linear (or polynomial) trends. A least-squares method is used to estimate the unknown coefficients. The technique, when applied to multi-decadal observations of a number representative stations, reveals complicated seasonal-diurnal patterns of trends in means and variances of meteorological variables. The technique can be applied to historical observations of meteorological stations with many changes in their times of observation.