Periodic components of water storage changes from GRACE and global hydrology models

We analyze spatiotemporal variations of surface mass anomalies induced by hydrological mass redistributions at the Earth\textquoterights surface. To this end, we use a suite of global hydrological models as well as products from the Gravity Recovery and Climate Experiment (GRACE) satellite mission. As a novelty we identify dominating periodic patterns that are not restricted to the fundamental annual frequency and its overtones, using a method that combines conventional empirical orthogonal functions (EOF) analysis with a determination of sine waves of arbitrary periods from the principal components. We assess the significance of the derived spectra in view of correlated errors of the GRACE data by means of a Monte Carlo technique. This allows us to create filtered GRACE time series including only the significant terms, which will serve for basin-specific calibration of hydrological models with respect to the dominant periodic water storage variations. The study reveals that besides dominating annual signals, semiannual (found only in a few basins), and also long-periodic waves in the range of 2.1 to 2.5 years contribute to periodic water storage variations. An interpretation and a preliminary explanation of these spectra is included. Comparisons of the spectra obtained from GRACE and global hydrological models exhibit in many river basins a systematic advance of the phases of annual terms of the hydrological models as compared to GRACE in the range of 1 to 6 weeks. This indicates deficiencies of the hydrological models with regard to runoff routing in the river network and/or water retention in lakes and wetlands.