*3.3. Multiscale/Multiphysics Capabilities*

This section describes how GCCOM is capable of handling complicated fluid behavior by means of two different methods. First, the multiple physical processes simulated inside the model, such as hydrostatic and non-hydrostatic pressure, sub-grid scale turbulence, thermodynamics and the density equations of state for seawater work together to capture complex processes. Second, the grid resolution we use will have a major impact on the detail level and richness of the physics we obtain. In this section, we present results of comparing the output of two different resolutions in the Seamount case to better illustrate the point of the multiple physics and multiple scales GCCOM can capture.

Figures 9 and 10 show the velocity flow along the horizontal axis. The images compare and contrast the High (left) vs. Low (right) grid resolution details for both side and bottom views of the domain. The rows of images show a series of zoomed-in details, represented by the rectangular boxes. The bars to the right of each image depict the scale values.

A side view comparison between the seamount test case of 3000 × 200 × 100 (high resolution, left side) and 1500 × 100 × 50 grid points (low resolution, right side) is shown in Figure 9 for the horizontal velocity, being forced on the right side of this figure. The high resolution has twice the resolution in each direction and eight times the number of grid points.This snapshot was taken at the mid-section of the domain after *t* = 6000 s of simulation. Each problem has been run with the same conditions and shows similar behavior.

In the top row of images, the kilometer scale is depicted along the horizontal axis. The image shows an accumulation of contours at the base of the seamount, and somewhat uniform velocities over the rest of the domain for both of the resolutions. The difference between the two plots resides in the density and locations of the contour lines. For the high resolution, there are significantly more contour lines around the seamount bathymetry than for the low resolution plots. Row 2 of Figure 9 shows a zoomed in detail of these structures. Here, the images show marked differences between the high and low resolution cases. In this frame, we see that the features developed in the high resolution panel (left) are not captured in the low (right) resolution case. At the same time, the richness of the higher resolution case can be explored further, as is shown in the bottom left panel. The increased magnification reveals a series of eddies, while in the lower resolution counterpart no special behavior is seen. These results demonstrate that the GCCOM model is capable of capturing more information as the grid resolution increases: an increase in the number of points translates to capturing richer and more complex phenomena across the domain; and the multiscale processes, ranging from kilometer to meter scale lengths.

The ability of the GCCOM model to capture both high and low resolution flow features is seen once again in Figure 10, where the view is from the bottom plane of the domain, where the velocity flows from East to West. Again, the flow is captured at *t* = 6000 s. We can see structures developing widely in the high resolution problem, while the low resolution only shows them happening in specific spots and in a broader distribution, but we see no sign of high resolution fluid structures in the rest of the domain. The middle rows in Figure 10 show an important difference in contour details: while the low resolution grid shows a structure that is similar to that of the high resolution grid, the high resolution grid captures the meandering waves of low velocity fields on the bottom of the seamount, something we could see if we were modeling the shape of a sandy sea bottom. The details and number of eddies behind the seamount peak are also richer in the high resolution grid, while only one broad eddy-like structure is seen in the low resolution case (bottom panels). The results shown in this section demonstrate that GCCOM is capable of capturing different types of phenomena including fluid flow, nonhydrostatic pressure and thermodynamics, over scales that range from 10<sup>0</sup> to 10<sup>3</sup> m, thus establishing that GCCOM is both a multiphysics and a multiscale model.

**Figure 9.** Figures depict zoomed-in details of the High (**left**) vs. Low (**right**) resolution side views of the stratified seamount experiment at *t* = 6000 s.

**Figure 10.** Figures depict zoomed-in details of the High (**left**) vs. Low (**right**) resolution bottom plane view for the stratified seamount experiment at *t* = 6000 s.
