utility

The utility test group is a general test group for ocean utilities that may create datasets for other test groups to use. It is partly designed to provide provenance for data processing that may be more complex than a single, short script.

extrap_woa

The class compass.ocean.tests.utility.extrap_woa.ExtrapWoa defines a test case for extrapolating WOA 2023 data into ice-shelf cavities and coastal regions, then into land, grounded ice and below bathymetry.

combine

The class compass.ocean.tests.utility.extrap_woa.Combine defines a step for combining the Annual and January temperature and salinity climatology data into a single file. We use the 30-year objectively analyzed climatology (1991-2020) data sets on a 0.25 degree lon-lat grid. The WOA data is provided only in the top 1500 m of the ocean in the monthly files, whereas it extends to 5500 m depth in the annual dataset. The test case uses the January data where it is available and the annual data when monthly data is not provided. Finally, the step converts in situ temperature to potential temperature.

remap_topography

The class compass.ocean.tests.utility.extrap_woa.RemapTopography defines a step for conservatively remapping a topography dataset (see Remapping Topography) from a very high resolution grid to the WOA 2023 0.25 degree grid. The topography and masks are then used to aid in the extrapolation process. Because of the very high resolution source grid, at least 360 cores are needed to make the mapping file in this step.

extrap_step

The class compass.ocean.tests.utility.extrap_woa.ExtrapStep defines a step for extrapolating the data into invalid regions. Because we will use the dataset to interpolate potential temperature and salinity at the centers of MPAS grid cells, we want there to be valid data not just where WOA 2023 has it defined but also in the cavities below Antarctic ice shelves, in a buffer region into land, and below the bathymetry. This way, the interpolation will be sure never to be contaminated with NaNs or other invalid values.

The ice-shelf cavities provide a particular challenge. No temperature or salinity data is available for them in WOA 2023 but they cover vast areas so it is not safe to simply extrapolate from the nearest point available in the WOA dataset. When we have done this in the past, properties deep in the Filchner-Ronne ice-shelf cavity come from the Bellinghshausen Sea, which is geographically close by but topographically disconnected from the cavity. To avoid this problem, we perform extrapolation in 4 steps.

  1. Extrapolate horizontally from WOA 2023 data into regions that the topography dataset indicates are ocean. This includes anywhere where the top of a WOA layer is above the bathymetry that is not covered the grounded antarctic ice sheet. This means that we extrapolate not only into ice-shelf cavities but also into the ice-shelves above them. Past experience has shown this to be a helpful approach.

  2. Extrapolate vertically downward from the surface to the seafloor. This fills in regions that are horizontally blocked by topography from any valid WOA data (e.g. the deep interiors of ice-shelf cavities) but which are still either in the ocean or in floating ice shelves.

  3. Extrapolate horizontally from the results of step 2 to fill in land, grounded ice sheet areas, and below the bathymetry. here, we will likely only ever use the data in a small halo around the “valid” ocean region when we interpolate it to MPAS-Ocean meshes but we fill in everywhere “just in case”.

  4. Finally, we extrapolate vertically downward one last time all the way to the bottom layer of the dataset. This won’t do anything unless there happen to be layers with no valid WOA data at all (which is not presently the case).

The resulting file is ready to be placed in compass’ initial condition database (see Downloading and symlinking input files for details on databases).