Mesh

Several ocean test groups use a set of common functionality for manipulating MPAS-Ocean meshes. This includes remapping topography datasets to the MPAS mesh, culling land from the mesh based on a series of masks.

Remapping topography

After building a base spherical mesh (see Spherical Meshes), the global ocean mesh test case includes a step for remapping topography data (bathymetry, ocean mask, land-ice draft, land-ice thickness, grounded and floating land-ice masks, etc.) to the MPAS mesh. This step is controlled by the following config options:

# config options related to remapping topography to an MPAS-Ocean mesh
[remap_topography]

# the name of the topography file in the bathymetry database
topo_filename = BedMachineAntarctica_v3_and_GEBCO_2023_0.0125_degree_20230831.nc

# variable names in topo_filename
lon_var = lon
lat_var = lat
bathymetry_var = bathymetry
ice_draft_var = ice_draft
ice_thickness_var = thickness
ice_frac_var = ice_mask
grounded_ice_frac_var = grounded_mask
ocean_frac_var = ocean_mask

# the description to include in metadata
description = Bathymetry is from GEBCO 2023, combined with BedMachine
              Antarctica v3 around Antarctica.

# the target and minimum number of MPI tasks to use in remapping
ntasks = 4096
min_tasks = 360

# remapping method {'bilinear', 'neareststod', 'conserve'}
method = conserve

The topography filename should be something from the ocean bathymetry database.

The various variable names are designed to provide flexibility in case a given topography file used different variable names tha the defaults provided above.

The description config option is passed on as part of the mesh metadata in global ocean files.

The target and minimum number of MPI tasks (ntasks and min_tasks, respectively) will depend on the resolution of the topography file. The default file is at 1/80 of a degree and typically requires at least 360 tasks to successfully remap the data even to relatively coarse MPAS meshes. Coarser bathymetry datasets can likely get away with far fewer MPI tasks. The method used for remapping (conserve by default) also makes a difference in how many tasks are required.

Culling land cells

The framework also includes a step for culling land from the MPAS mesh, including enforcing a series of critical passages (transects that must be ocean, such as narrow channels) and critical land blockages (transects that must be land, such as thin peninsulas). The config options that can be used to control this step are:

# options for spherical meshes
[spherical_mesh]

## config options related to the step for culling land from the mesh
# number of cores to use
cull_mesh_cpus_per_task = 128
# minimum of cores, below which the step fails
cull_mesh_min_cpus_per_task = 1

To create various land masks, the culling step uses python multiprocessing. The target and minimum number of processes are controlled by cull_mesh_cpus_per_task and cull_mesh_min_cpus_per_task, respectively.