Time series of precipitation from three satellite products (TRMM 3B42, CMORPH, and PERSIANN) and two reanalysis models (the NCEP/DOE Reanalysis-2 and ECPC Seasonal Forecast Model) are examined for their behavior at different frequencies using both harmonic and Fourier approaches. Power spectra generated by fast-Fourier transforms are divided into wide bands comprising the low-frequency (period greater than 30 days), synoptic (period between 2 – 30 days), and high-frequency (period less than 2 days) components. Low-frequency variability is generally the weakest component, but is amplified over the Inter-Tropical Convergence Zone and monsoonal regions. Land areas feature enhanced high-frequency variability and reduced synoptic variability in comparison to the surrounding oceans. The diurnal and semidiurnal harmonics explain very little of the variance, suggesting that the harmonic approach does a poor job of capturing the high-frequency variability of the datasets. Inter-satellite and inter-model differences also indicate biases of the precipitation product algorithms and convective parameterizations, including a strong bias toward low-frequency variability in the Relaxed Arakawa-Schubert parameterization employed by the ECPC Seasonal Forecast Model.