Concept
Audience
Early career researchers new to working with large climate datasets, with
some experience with Python programming
Goals
- Able to access and analyse large datasets such as CMIP5 efficiently
- Able to set correct metadata on output files for publication
Length
Half day
Summative Assessment:
- Calculate NINO34 index on official ACCESS 1.3 dataset, storing output as CF-netCDF
Covers concepts:
- Catalogue
- THREDDS
- Syphon
- Xarray
- Dask
- CF-Metadata
- Compare a model’s output with re-analysis, plotting result
Covers concepts:
- Search for datasets on the Catalogue
- Explore available datasets, THREDDS functionality
- Get URLs for a data run with Syphon
- Collect list of files holding the ocean surface temperature
- Load a single file with Xarray & inspect contents
- See variables, attributes. Compare with
ncdump
- Average & calculate anomaly over NINO34 area
- Subset the data, explore chunking options
- Explain Dask, memory limits
- Load multiple files with chunking
- Compare loops vs whole-array operations
- Use cdo/nco to calculate seasonal average
- Importance of metadata, history attribute
- Quick plots with ncview
- Prepare publication, DOIs, ANDS, Orchid
- http://climate-cms.unsw.wikispaces.net/Data+publishing+guidelines
- Data citation, reproducible data
- Load a model data cube with Iris
- Different level types, aux dimensions
- Plot the difference between model and ERA-Interim surface temp with Cartopy