The performance of a z-level ocean model in modeling the global tide

Abstract

The performance of a z-level ocean model, the Modular Ocean Model Version 4 (MOM4), is evaluated in terms of simulating the global tide with different horizontal resolutions commonly used by climate models. The performance using various sets of model topography is evaluated. The results show that the optimum filter radius can improve the simulated co-tidal phase and that better topography quality can lead to smaller rootmean square (RMS) error in simulated tides. Sensitivity experiments are conducted to test the impact of spatial resolutions. It is shown that the model results are sensitive to horizontal resolutions. The calculated absolute mean errors of the co-tidal phase show that simulations with horizontal resolutions of 0.5° and 0.25° have about 35.5% higher performance compared that with 1° model resolution. An internal tide drag parameterization is adopted to reduce large system errors in the tidal amplitude. The RMS error of the best tuned 0.25° model compared with the satellite-altimetry-constrained model TPXO7.2 is 8.5 cm for M2. The tidal energy fluxes of M2 and K1 are calculated and their patterns are in good agreement with those from the TPXO7.2. The correlation coefficients of the tidal energy fluxes can be used as an important index to evaluate a model skill.