Skip to main content
U.S. flag

An official website of the United States government

A Testbed to Measure Spurious Mixing in MPAS-Ocean

Presentation Date
Tuesday, May 13, 2014 at 5:00pm
Authors

Author

Abstract

Spurious numerical mixing is an important consideration in the design of algorithms and vertical coordinates in ocean-climate models. Proper mixing is critical to set the correct water mass properties, which then influence circulation patterns. This is particularly true of overflow regions in the North Atlantic and the resulting meridional overturning circulation. Here we present a series of standard tests to quantify spurious mixing, ranging from a simple two-fluid lock exchange to real-world global ocean simulations (Ilicak et al. 2012). The Model for Prediction Acros Scales-Ocean (MPAS-Ocean) is compared directly to three other ocean-climate models. The test cases validate that MPAS-Ocean dynamics and mixing behave as expected. MPAS-Ocean generally has less spurious mixing than other ocean models. Within a given model, the mixing testbed is useful to compare choices of vertical coordinates. The MPAS-Ocean vertical coordinate is structured arbitrary Lagrangian-Eulerian (ALE). When fully Eulerian, MPAS-Ocean is a z-level model with fixed thicknesses; when fully Lagrangian, there is no vertical transport between layers, and layers expand and contract like an isopycnal ocean model. In between are many additional options, such as z-star where layers expand in proportion to the sea surface height, and sigma, where coordinates are terrain-following. Sigma coordinates dramatically reduce spurious mixing in the idealized overflow test case. MPAS-Ocean includes a ''z-tilde'' option (Leclair and Madec, 2011), where the low frequency thickness oscillations are filtered out. Then high-frequency thickness oscillations, such as internal gravity waves, are treated in a Lagrangian manner. The mixing testbed cases show that z-tilde reduces spurious vertical mixing and preserves water mass properties. MPAS-Ocean, developed at LANL, was first released in June 2013 (see http://mpas-dev.github.io/). It has been tested on high-resolution global quasi-uniform and variable-resolution domains with realistic topography (Ringler et al., 2013). MPAS-Ocean now includes most of the features of a full ocean-climate model. The 15km global simulation achieves two simulated years per wall-clock day on 3000 processors, identical to 0.1-degree POP, a rectangular-grid model. MPAS-Ocean is currently being coupled to the Community Earth System Model (CESM), so that testing may proceed with coupled atmosphere, ice, and land surface components. Ilicak, M., Adcroft, A.J., Griffes, S.M., Hallberg, R.W., 2012: Spurious dianeutral mixing and the role of momentum closure. Ocean Modelling, 45, 37-58 Leclair, M. and G. Madec, 2011: z-tilde Coordinate, an Arbitrary Lagrangian-Eulerian coordinate separating high and low frequency motions. Ocean Modelling, 37,

Funding Program Area(s)