*********************************************************** ******** Introduction ***********************************************************************

This webiste contains a live version of the EOLDAS documentation. EOLDAS stands for Earth Observation Land Data Assimilation System, and in here we refer to an implementation of a Data Assimilation (DA) prototype to be used to monitor the land surface in the optical domain. The code is fairly generic, and can be used to implement a number of different variational DA schemes (although we mostly have used weak constraints)

This software was developed

Note

Partners, project number, thanks to etc.

Installation¶

Requirements¶

EOLDAS has mostly been written in Python, with some radiative transfer codes written in Fortran and made available to the main Python library. The requirements for the package are

Python (versions > 2.5 and < 3.0)
Scipy
Numpy
Matplotlib
gfortran
OpenOpt (optional)
git (version control system. Optional)

The first few packages are usually available in modern Linux distributions, as well as in MacOsX. In some cases, the use of the Enthought Python Distribution to install a whole Python ecosystem might be a worthy option. Similar efforts for Windows might be found in Python xy.

OpenOpt is an optimisation suite, and is optional, although its use is highly recommended. Using pip or easy_install, the Python package installers, it can be installed as

easy_install -U openopt

pip install openopt

Note that if you do not have root/superuser access, you can install it for your user as

easy_install --user -U openopt

pip install --user openopt

Installation¶

The eoldas distribution can be installed using the same approach as OpenOpt:

easy_install --user -U eoldas

or:

pip install --user eoldas

This will install as an user, ignore the --user option to do a system-wide install. Note that executables will be installed too, and the installation executables path will be installation dependent. Note that this command will also install the semidiscrete radiative transfer code and compile it (many warning will show up, but they are safe to ignore).

Installing from source¶

If you plan to do development on EOLDAS, we strongly encourage you to fork the EOLDAS repository on github, and feel free to submit patches, suggestions, etc. You can also download a snapshot of the code from github. Once you have unpacked the tarball, you can install it using the following command

python setup.py install

(as before, if you don’t have super user priviledges, add the --user flag to the previous command).

Installing the user’s guide¶

The user’s guide is available on github.com. Although you can just download an archive with the examples, data and source of this guide, we encourage you to fork the eoldas_release git repository, and to work on it and submit bug reports, fixes etc.

Getting help & Collaborating on improving EOLDAS¶

EOLDAS is a complex and varied tool that can help solve many problems where Gaussian statistics are a good assumption in variational problems. We would like users to use github as a way to collaborate in fixing bugs, spreading what you can do with EOLDAS, and to put together tutorials.

There are two main ways for collaboration in either the user’s guide or the main EOLDAS package: one is to make a pull request after you have forked the relevant repository (eoldas or eoldas_release). A second avenue is to use the Wiki. The user’s guide has Wiki page, and so does the main eoldas repostiory.