Extracting timeseries from many models from point locations
Developer: Gen Tolhurst (E: [email protected])
Supervised by: Mitch Black (E: [email protected])
Funded by: ACS
The bulk of the processing is done using this parallelised bash script data_extraction.sh
Follow data_extraction.ipynb for the process of extracting time series from many point locations.
The process constructs a csv for each location with daily tasmax, tasmin, and tasmean values from 13 models. This results in 3 x 13 columns of temperature data. The dataframe will look something like this:
UCL_2021_AUST_with_location_top_33.csv is an example of a list of point locations that were requested for time series extraction. From that csv file, we simplified the file to include only the location code, lon, and lat UCL_2021_AUST_top_33_coords_only.csv.
This notebook catalogues the ACS bias-adjusted data acs_catalogue_data.ipynb. This method leans heavily on code Gen previously developed at https://git.nci.org.au/pfd548/bom-climate-change-variability-and-extreme-toolbox/-/blob/master/catalog_builder.py.
Xarray can extract smaller time series using methods shown in data_extraction_xr_example.ipynb