Comparison of mass anomalies from the Gravity Recovery and Climate Experiment (GRACE) to in-situ ocean bottom pressure (OBP) data is complicated by the fact that GRACE estimates are representative for phenomena with scales of hundreds of km, while OBP measurements are pointwise by nature. Simulations with the Finite Element Sea-ice Ocean Model (FESOM) indicate that on a monthly time scale OBP anomalies are coherent over large areas of complex geometry. A new filtering method for GRACE-derived ocean mass anomalies is obtained by applying these coherence patterns on a regular grid. Time series from OBP recorders all over the world ocean are used for validation. Compared to Gaussian filtering, the pattern-filtering method improves the correlation between GRACE retrievals and in-situ OBP data (max. correlation now 0.87). Land leakage effects are eliminated almost completely. Finer scales can be resolved such that highly variable currents like the Gulf Stream can now be identified.