Ocean tides redistribute mass at high temporal frequencies. Satellite missions that aim to observe medium to low frequency mass variations need to take into account this rapidly varying mass signal. Correcting for the effects of ocean tides by means of imperfect models might hamper the observation of other temporal gravity field signals of interest. This paper explores different methods for mitigating aliasing errors for the specific example of observing mass variations due to land hydrology, including temporal filtering of time-series of gravity solutions, spatial smoothing and the use of satellite constellations. For this purpose, an Earth System Model (ESM) was constructed, which included state-of-the-art time varying components for ocean, atmosphere, solid Earth, hydrology, ice-sheets and ocean tides. Using the ESM, we simulated the retrieval of the hydrologically driven gravity field changes using a number of different satellite constellations.We find that (1) the ocean tide aliasing strongly depends on the satellite constellation, the choice of orbital parameters and the length of the data span; (2) the aliasing effect manifests itself differently for different geographical regions; (3) the aliasing causes a peculiar striping pattern along the ground track of the satellite orbits; (4) optimizing the choice of orbital parameters of a single GRACE-type tandem can be more effective at reducing the aliasing of ocean tide model errors than flying more tandems. Finally, we corroborate the experiences with GRACE data analysis that appropriate post-processing techniques can significantly improve the quality of the retrieved gravity changes. © 2010 The Authors Journal compilation © 2010 RAS.