search

john-science / python_for_scientists

Python Open Courseware for Scientists and Engineers
GNU General Public License v3.0
68 stars 40 forks source link

Is NetCDF4 sufficient? #32

Closed john-science closed 9 years ago

john-science commented 9 years ago

Jeremy's group has concerns about NetCDF4 being sufficient. Particularly for IOAPI-based NetCDF files. Here is a list of things they need to do:

  1. average multiple files
  2. convert lon/lat coordinates for monitors to I/J coordinates and extract the corresponding model data
  3. sum multiple model species (and multiple by a scalar to convert units)
  4. calculate running average mixing ratios (e.g., for calculating 8-hr average ozone)
  5. calculate county, basin, district total emissions from the model-ready emissions

We generally don't do anything too crazy to the netcdf files. The real key is to be able to open files, update those files, create new files, read/write from files, combine files, etc. Pretty basic stuff, but from there folks can write any program they need.

john-science commented 9 years ago

James suggested this as an alternative: https://github.com/barronh/pseudonetcdf

john-science commented 9 years ago

I imagine NetCDF4 is plenty sufficient. But here are some helpful links:

NetCDF docs: http://www.unidata.ucar.edu/software/netcdf/docs/ IOAPI requirements: https://www.cmascenter.org/ioapi/documentation/3.1/html/REQUIREMENTS.html

john-science commented 9 years ago

I have been looking into it, and I don't see any problems with IOAPI using netCDF4. And since that is a much better supported and widely-used library, I'll just go with that.