Integrating models and observations

OpenDA is an open interface standard for (and free implementation of) a set of tools to quickly implement data-assimilation and calibration for arbitrary numerical models. OpenDA wants to stimulate the use of data-assimilation and calibration by lowering the implementation costs and enhancing the exchange of software among researchers and end-users.

A model that conforms to the OpenDA standard can use all the tools that are available in OpenDA. This allows experimentation with data-assimilation/calibration methods without the need for extensive programming. Reversely, developers of data-assimilation/calibration software that make their implementations compatible with the OpenDA interface will make their new methods usable for all OpenDA users (either for free or on a commercial basis).

OpenDA has been designed for high performance. Hence, even large-scale models can use it. Also, OpenDA allows users to optimize the interaction between their model and the data-assimilation/calibration methods. Hence, data-assimilation with OpenDA can be as efficient as with custom-made implementations of data-assimilation methods.

OpenDA is an Open Source project. Contributions are welcome from anyone wishing to participate in the further development of the OpenDA toolset.

Release of version 1.0

The release of version 1.0 of OpenDA has been publicly announced on May 10, 2010 during the Jonsmod conference in Delft, the Netherlands. Information related to the release can be found here. At this moment, the Windows binary version is available through the OpenDA website. As of June 20, 2010, the complete release is expected to be available through sourceforge.net.

Features of OpenDA

Currently available methods

data-assimilation methods
- Ensemble method
- Reduced rank square root method
- Ensemble Square root method
- COFFEE method
Parameter estimation (calibration) methods:
- Simplex method
- Conjugate gradient method
- LBFGS method