VACUMM provides generic and specialized tools for the validation of ocean models, and more especially the MARS model from IFREMER. The heart of VACUMM is a library written mainly in the Python language, whose core can be used for the preprocessing and the postprocessing of oceanic and atmospheric data coming from models or observations. The library for instance also has specialized modules for managing outputs from models and making advanced diagnostics.

>>> from vcmq import *
>>> sst = DS(data_sample(''), 'mars').get_sst()
>>> map2(sst)


  • A huge documentation with a gallery, a lot of examples and the complete API:
  • Full UV-CDAT support and extensions.
  • Matplotlib/basemap graphics with advanced plotting objects like geographical mapping tools.
  • Numerous utilities for manipulating and converting time data.
  • Regridding and interpolation of random or gridded data, in 1D or 2D, with curvilinear grid support.
  • Helper routines for inspecting and reading NetCDF objects in single or multiple file datasets.
  • Generic and specialized 1D and 2D filters working on masked variables.
  • Specialized physical and numerical diagnostics, like dynamics, thermodynamics, spectral analyses, tides, etc.
  • Support and extension of CF conventions for searching or formatting variables.
  • Miscellaneous location utilities such as readers of sigma coordinates for ocean models, or Arakawa grid converters.
  • High level generic interface for reading and post-processing NetCDF data from standard or known dataset, such as model outputs or satellite products.
  • Statistical accumulator for large datasets.
  • Interfaces for working with random and gridded bathymetries, and with shorelines.
  • Utilities for working with masks and selection of spatial data.
  • Utilities for working with input and output remote files.
  • Advanced logging classes.
  • Extensions to sphinx for Fortran and Python.
  • A collection of scripts for some diagnostics.


Mandatory: CDAT (or more specifically cdms2, cdutil, genutil from CDAT, and matplotlib, basemap), configobj.

Optional: seawater, PIL, pytz, paramiko, xlwt, sphinx-fortran, cmocean.


To download VACUMM sources, please go to this page:


From sources:

$ python install

Using conda:

$ conda install -c vacumm -c conda-forge -c cdat  vacumm

For more information, please go to this:

Release notes

Release notes for each version are available here:


The documentation is available here:


VACUMM is under the CeCiLL license, which is compatible with well knwon GPL license.


You can submit issues and pull requests from the GitHub site.

Stephane Raynaud (raynaud (at), Guillaume Charria (Guillaume.Charria (at)

See the contact page: