This branch is for taking upstream CCMP data and reducing it to weekly averages in an intermediate set of netcdf files; see branch db-population for writing these weekly averages to mongodb.
- this Argovis product based on the 6h wind measurements at https://www.remss.com/measurements/ccmp/
- build and push container image described in
Dockerfile-ftpasargovis/ccmp:download - put REMSS FTP credentials in
pod-template.yaml - edit and run
dl.shto download data as desired. - a large openshift volume named
ccmpis expected to be available.
- weekly averages aligned to match https://psl.noaa.gov/data/gridded/data.noaa.oisst.v2.html
- build and push image described in
Dockerfileasargovis/ccmp:reduceand run viapod-reduce.yamlafter choosing the appropriate year. Resource intensive, takes about 14h on openshift.
- build and push image described in
Dockerfile-proofreadasargovis/ccmp:reduce-proofreadand run viapod-prrofread.yaml - Randomly checks for consistency between upstream and reduced data until interrupted.
- Use
dump.pyfrom the same container to manually inspect the upstream values for a given lat/lon in a given upstream file.