This repository is intended to store, where appropriate, the many scripts and tools used by the land community within Australia as part of their research for ancillary data creation. This includes processes such as re-gridding, aggregation and disaggregation, statistical analyses and reductions, resampling, reformatting of datasets for ingestion by either CABLE or JULES. The goal is to use these community-shared resources to converge towards a shared infrastructure to create ancillary data.
This repository was first created for the working bee day at the 2024 Land Surface Model Working Group workshop.
Many groups within the land community are using similar workflows and have similar requirements. By collecting and cataloguing the tools created by the community in one place, we achieve the following objectives:
- Reduce duplicated work by reusing existing functionality
- Ensure comparability between scientific runs made by different groups by using unified methods of interpolation, aggregation etc.
- Results are reproducible, as we have a history of any changes made to the tools
Place your script or tool in the relevant directory within the repository and include a README. The README should be named according to the name of your script e.g. if your script is called interpolate_met_data.py, then your README should be named interpolate_met_data_README.txt.
The README should have three headings: Purpose, Usage, Method. Please be verbose in the descriptions. For example:
BAD: Interpolates met data
GOOD: Interpolates the CRU-JRA met data to the desired resolution and writes the output to NetCDF in a format readable by CABLE.
BAD: interpolate_met_data.py -r lat_resolution,lon_resolution -o output.nc -b lat_min,lat_max,lon_min,lon_max
GOOD: Call from the command line via interpolate_met_data.py -r lat_resolution,lon_resolution -o output.nc -b lat_min,lat_max,lon_min,lon_max, where:
- "-r lat_resolution,lon_resolution" are the grid spacings (required).
- "-o output.nc" specifies the name of the output file to write to (required).
- "-b lat_min,lat_max,lon_min,lon_max" specify the domain to interpolate (optional). If no bounds are specified, then the original domain is used.
BAD: uses linear interpolation
GOOD: Reads the dataset using xarray and uses scipy's LinearNDInterpolator to perform the interpolation.
The repository is currently structured to reflect the categories set on the working day i.e. ancillaries and meteorological forcing. The structure for this repository (or it's successor) will be open to discussion.
