Accessing NetCDF4 Data in Python

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Monday, 5 January 2015
Ward Fisher, UCAR/NCAR, Boulder, CO

In the past few years, Python has increasingly become an important tool for scientific investigation. Before Python, scientists might be required to master low-level computer languages like ‘C' or ‘Fortran' in order to analyze their data. Because much of the complexity found in these traditional programming languages has been abstracted away, Python is highly accessible for non-computational scientists. When coupled with packages like ‘numpy', ‘scypi', and ‘matplotlib', Python enables scientists to make sense of their data and to carry out complex data analysis in a practical manner.

The netCDF file format is broadly used in the atmospheric sciences for data archival. Until recently, scientists using the netcdf4 enhanced data model did not enjoy a reliable Python-native solution when working with this data. They would instead need to employ cumbersome and complex external solutions in order to import their data into Python for analysis.

The netcdf4-python package, created by Jeff Whitaker at the NOAA Earth System Research Laboratory, is a Python/numpy interface for the netCDF4 library. This package has been embraced by Unidata as a tool for working with netCDF4 in Python, and with the support of Jeff Whitaker, netcdf4-python has been included under the umbrella of the Unidata Github repository.

With netcdf4-python it is possible for scientists to carry out their data analysis in python using data stored in the netCDF4 enhanced data model. This work will demonstrate how netcdf4-python can be used to read, subset, analyze, and visualize data stored in the netCDF4 format.