Implementation of GEBCO cookbook remove-restore and other cleaning of topography/bathymetry. Uses pyGMT.

The non-python dependencies are:

• gdal
• GMT

Some dependencies of the package are a bit fiddly to install, especially on Windows. We recommend using a conda environment, and installing the more difficult dependencies through conda-forge, by running the commands below in sequence:

1. The proceeding steps assume your current working directory is your local copy of the repository. A mixture of conda and pip is used to install the dependencies (due to the complex dependencies gdal and rasterio). The conda dependencies are listed in environment.yml which will be automatically used by conda when creating a new environment:

conda env create

3. This will create a new conda environment named pycascadia which should be activated with:

conda activate pycascadia

4. Install pyCascadia. This will also automatically install the remaining dependencies (a list of which can be found in setup.cfg).

pip install pycascadia

5. Once you've done your work with pyCascadia, you may want to deactivate the environment and return to your base python environment. You can do so by running:

conda deactivate

Before using any part of pyCascadia, make sure you've got its conda environment activated (change pycascadia to the name of your environment if you've named it differently when following the installation instructions).

conda activate pycascadia

pyCascadia provides the remove restore algorithm as a command line tool called remove-restore, e.g.

remove-restore --base gebco_base_grid.nc higher_res_grid.tiff --output merged_grid.nc

Input base and source grids are accepted in both GeoTiff or NetCDF formats. It is possible to provide more than one source grid, e.g with three source grids one would call:

remove-restore --base gebco_base_grid.nc higher_res_grid1.tiff higher_res_grid2.tiff higher_res_grid3 --output merged_grid.nc

For more details on these and other input arguments of remove-restore, run

remove-restore -h

on your command line.

For more fine-grained control of the remove-restore functionality, there is an example Jupyter notebook under ./notebooks/. Note that to use this, you will need to pip install jupyter in the conda environment, too.

Details of the full API can be found in the documentation.

remove-restore expects a single variable per netCDF file, which is not always the case (see issue #59). The delete-variable tool can be used to remove a single variable from a netCDF file in the following way:

delete-variable --input <input.nc> --output <output.nc> --variable <varname>

The script close_boundary.py provides a way to replace values along the domain boundary. This can be used to create "closed" shape files, that are necessary for meshing in OceanMesh2D, in combination with the GDAL command gdal_contour. The close_boundary.py file can be found in the scripts/ subfolder of your local copy of this GitHub repository, so you should navigate there to use them (or replace the relative path in the commands below, e.g. python close_boundary.py -> python ./scripts/close_boundary.py).

cd scripts/

Instructions for the main use case - creating shapefiles that consist of closed contours at a certain value $z$ and on the domain boundary, where the values are $>z$:

1. Generate a new netCDF file with all domain boundaries $>z$ set to a value slightly lower than $z$, i.e. $z-epsilon$:

python close_boundary.py --value <z-epsilon> --input <input.nc> --output <output.nc>

The reason for using a slightly lower value here is because if the value is exactly equal to $z$, gdal_contour does not interpret these locations as a contour line at $z$.

1. Run gdal_contour on the output.nc file of the previous step to create the final closed.shp shape file:

gdal_contour -fl <z> <output.nc> <closed.shp>

In our wider Cascadia pipeline, we use contours based on the proximity-to-the-coast map, as these are smoother and avoid discontinuities in the contour lines, which would prevent us from meshing.