Food Shocks Cascade

Food Shocks Cascade (FSC) is executable through MaaS using a Docker container which Jataware generated in collaboration with Columbia University. However, this Docker container produces a complex set of outputs which was not relevant for the January Experiment. Instead, Columbia provided a pre-run set of outputs which were in a simplified format driven by the needs of the Experiment use case. So, execution of FSC has been disabled in MaaS for the time being.

Currently, FSC is limited to wheat only.

Format

FSC outputs are all .csv files. It’s inputs are also .csv files. The pre-run output for FSC is available here.

Resolution

FSC output is at the country-level and is annual.

Processing

The MaaS controller for FSC is called fsc.py. This is currently disabled in favor of the pre-run output provided by Columbia. That output was processed with the FSC_processing.py script.

Issues/Lessons Learned

• FSC is a relatively lightweight R based model. It is easy to execute but currently its executable Docker container does not output data in the same format as has been provided by Columbia for its pre-run output. Once Columbia updates the Docker container to produce output in this format it should be trivial to pull the updated container and turn back on FSC.
• FSC was an example of a model that required thresholding for visualization purposes. The “shock severity” output variable was based on thresholds set by Columbia; these were adjusted after Columbia determined that the thresholds they set did not provide sufficient visual differentiation between severity levels.