National Wind Risk Assessment
This repo collects the code developed as part of Geoscience Australia's contribution to the National Wind Risk Assessment (NWRA), a component of the National Climate Risk Assessment (NCRA; 2023-24).
Regional wind hazard
The regional hazard data is generated in ArcGIS Pro, using geoprocessing functions to convert shapefiles to rasters.
The Regional wind hazard provenance.ipynb notebook is used to generate provenance information for the regional hazard data. Running the code in the notebook will generate the provenance file.
applyMultpliers.cmd is a Windows command file to run the applyMultipliers.py script. This will combine the regional wind hazard data with the M3_max site exposure data to generate local wind hazard data. While the Windows command file is the main way to run this script, there's nothing to stop you running the python script on a *NIX machine. Just ensure the conda environment is properly set up (I haven't added a conda environment file yet CA 2023-12-05)
Vulnerability
Validate vulnerability file.ipynb is used to ensure the XML file that holds the vulnerability files is correctly formed NRML v0.5 data.
Vulnerability model NRML file provenance.ipynb generates the provenance information for the vulnerability model file.
Impact calculations
run_hazimp_batch.sh is submitted to the job queue on gadi.nci.org.au to run a collection of HazImp simulations in parallel. This script bulds a list of commands based on the available wind hazard layers in the scratch partition, and then calls run_hazimp.sh to execute HazImp for each wind hazard layer.
run_hazimp_batch.sh is set up to run a single region (i.e. State/Territory), which is specified when submitting the job.
To run, edit the template configuration file hazimp_template.yaml. RPX will be replaced with strings like RP100, while STATE will be replaced with the string value for the region ("WA", "NT", etc.). In most cases, the output folders (and any required parent directories) will be created in run_hazimp.sh.
Submit the job script to the queue, specifying the region::
qsub -v STATE=QLD run_hazimp_batch.sh
There is no checking done on the regions - if you try "NSZ", then the job will fail when it tries to actually run HazImp and finds there's no exposure file that starts with "NSZ". Currently the list of regions is "WA", "SA", "NT", "TAS", "VIC", "NSW", "ACT" and "QLD". "OT" won't work, because there is no hazard data for the other territories.
The resources specified in run_hazimp_batch.sh should be sufficient for all regions. Check the bigger ones ("NSW", "VIC") - increasing memory should do the trick.
Copyright
This work © 2023 by Commonwealth of Australia (Geoscience Australia) is licensed under CC BY 4.0 International
Contacts
Craig Arthur, craig.arthur@ga.gov.au