= PDAF - the Parallel Data Assimilation Framework =

{{{
#!html
<table align="right" width=300 cellpadding=10><tr><td>

<table style="margin: 0 0 0 0; border: 1px outset #ddc; background: #ccf;" cellpadding=5><tr><td align="center">PDAF is developed, <br>hosted and maintained by the<br><a href="https://www.awi.de/en/science/special-groups/scientific-computing/data-assimilation.html" target="_blank">Data Assimilation Team in the Computing Center of the Alfred Wegener Institute</a>.</td></tr></table>

</td></tr><tr><td>

<table style="margin: 0 0 0 0; border: 1px outset #ddc; background: #ccf;" cellpadding=5><tr><th align="center"><b> 8th WMO Symposium on Data Assimilation & 12th International Symposium on Data Assimilation</b></th></tr>
<tr><td>From August 31 - September 4 in Nanjing China. Applications and developments of PDAF are welcome contributions. For more information see the <a target="_blank" href="https://da2026.cms1924.org">Conference website</a>. </td></tr>
</table>

</td></tr><tr><td>

<table style="margin: 0 0 0 0; border: 1px outset #ddc; background: #ccf;" cellpadding=5><tr><th align="center"><b>ISDA-Online</b></th></tr>
<tr><td>We co-organize the International Symposium on Data Assimilation - Online. For more information see the <a target="_blank" href="https://isda-online.univie.ac.at/">ISDA-Online website</a></td></tr>
</table>

</td></tr><tr><td>

<table align="right" style="border: 1px outset #ddc; background: #ccf;" cellpadding=5><tr><th align="center"><b>News</b></th></tr>
<tr><td><u>April 16, 2026</u><br>
<b>Release of Version 3.1</b> - Feature release with new functionality (<a href="/trac/wiki/ReleaseNotes">release notes</a>)</td></tr>
<tr><td><u>November 5, 2025</u><br>
<b>Paper on pyPDAF</b> - The paper on pyPDAF is finally published by GMD: <a href="https://doi.org/10.5194/gmd-18-8235-2025">doi:10.5194/gmd-18-8235-2025</a>. </td></tr>
<tr><td><u>September 1, 2025</u><br>
<b>Release of Version 3.0</b> - Completing the major code revision and modernization including new functionality (<a href="/trac/wiki/ReleaseNotes">release notes</a>)</td></tr>
<!--<tr><td><u>May 30, 2025</u><br>
<b>Release of Version 3.0beta</b> - Major code revision and modernization including new functionality (<a href="/trac/wiki/ReleaseNotes">release notes</a>)</td></tr>//-->
<tr><td><u>April 3, 2025</u><br>
<b>New model coupling</b> to PDAF: application to <b>SCHISM ocean model</b> by Yu et al. published in Ocean Modling, <a href="https://doi.org/10.1016/j.ocemod.2025.102546">doi:10.1016/j.ocemod.2025.102546</a>
<tr><td><u>December 9, 2024</u><br>
<a href="https://pdaf.awi.de/trac/wiki/pyPDAF"><b>pyPDAF</b></a> allows to use PDAF purely with Python. PyPDAF is developed by Yumeng Chen, University of Reading. See the <a href="https://pdaf.awi.de/trac/wiki/pyPDAF">Page on pyPDAF</a> for further information. </td></tr>
<!--<tr><td><u>July 11, 2023</u><br>
New model coupling to run PDAF with the <b>NEMO ocean model</b>. Available on Github: <a target="_blank" href="https://github.com/PDAF/NEMO-PDAF">github.com/PDAF/NEMO-PDAF</a></td></tr>//-->
<!--<tr><td><u>April 27-May 2, 2025</u><br>
<b><a href="/trac/wiki/EGU2025">PDAF at the EGU General Assembly 2025</a></b> There will be several presentations of research in which PDAF is used. In addition, there will be a <b>short course</b> on data assimilation on Wednesday.</td></tr>//-->
<!--</td></tr><tr><td><u>July 11, 2023</u><br>
PDAF is now also available on Github: <b><a target="_blank" href="https://github.com/PDAF/PDAF">github.com/PDAF/PDAF</a></b>. Also see <a target="_blank" href="https://doi.org/10.5281/zenodo.7861829">Zenodo for release packages</a>.</td></tr>
<tr><td><u>April 24-28, 2023</u><br>//-->
<!--<tr><td><u>November 19, 2020</u><br>
Job announcement: <a href="https://recruitingapp-5442.de.umantis.com/Vacancies/698/Description/2"><b>Postdoc position on Biogeochemical Modeling and Data Assimilation</b></a> (closing date December 16, 2020)</td></tr>//-->
</table>

</td></tr></table>
}}}


The Parallel Data Assimilation Framework - PDAF - is a software environment for data assimilation. PDAF simplifies the implementation of data assimilation systems with existing numerical models. With this, users can obtain a data assimilation system with less work and can focus on applying data assimilation.

PDAF provides fully implemented and optimized data assimilation algorithms, in particular ensemble-based Kalman filters like LETKF, EnKF and LESTKF as well as nonlinear filters and variational methods (3D-Var and 3D ensemble Var). It allows users to easily test different assimilation algorithms and observations. PDAF is optimized for the application with large-scale models that usually run on high-performance computers and is applicable for operational applications. However, it is also well suited for smaller models and even toy models and can be used to teach data assimilation.

PDAF provides a standardized interface that separates the numerical model from the assimilation routines and the observation handling. This allows independent further development of the assimilation methods and the model, as well as the implementation of further observations. PDAF makes new algorithmic developments readily available through the interface such that they can be immediately applied with existing implementations. The PDAF release provides small models for easy testing of algorithmic developments and for teaching data assimilation. There are also existing coupling codes for different models of Earth system components.

PDAF is a community open-source project. Its functionality will be further extended by input from research projects. In addition, users are welcome to contribute to the further enhancement of PDAF, e.g., by contributing additional assimilation methods or coupling routines for different numerical models, or observation operators. 


{{{
#!html
<table class="wiki">
<tr><td><b>Information on PDAF V3.x for users of PDAF2:</b><br>
With PDAF V3.0 the PDAF code was modernized and not fully compatible with previous releases of PDAF V2.x. There are some small adaptions required for existing codes, but there is also an ample amount of new functionality and a new universal calling interface:
<ul>
<li><a href="https://pdaf.awi.de/trac/wiki/PortingToPDAF3">Information on adaptions for using PDAF V3.x with previous implementations using PDAF V2.x</a></li>
<li><a href="https://pdaf.awi.de/trac/wiki/PDAF3_new_functionality">New functionality in PDAF V3.0</a></li>
<!--<li><a href="https://pdaf.awi.de/trac/wiki/PDAF3_interface">The new universal PDAF3 assimilation interface</a></li>
<li><a href="https://pdaf.awi.de/files/PDAF3release_webinar_2025-06-11.pdf">Slides from PDAF3 release webinar on June 11, 2025</li>//-->
</ul>
</td></tr></table>
}}}

=== Information, documentation, and tutorials ===


{{{
#!html
<table class="wiki">
<tr><td>The documentation has been revised for PDAF V3. However, please excuse if some links do not yet work properly. 
</td></tr></table>
}}}

 * [FeaturesofPdaf Features and requirements of PDAF]
 * [wiki:ModelsConnectedToPDAF List of models connected to PDAF]
 * [GeneralImplementationConcept Implementation Concept of PDAF]
{{{
#!html
<li><a class="wiki" href="/register/index.php">Software download and email list</a></li>
}}}
 * [wiki:pyPDAF pyPDAF, using PDAF with Python]
 * Basic Documentation
   * [wiki:FirstSteps First Steps with PDAF]
   * [wiki:Lorenz_96_model Assimilation with the Lorenz-96 model and PDAF]
   * [wiki:Lorenz_63_model Assimilation with the Lorenz-63 model and PDAF]
   * [wiki:SoftwarePackage Description of the software package]
   * [wiki:PdafTutorial Implementation tutorials]
   * [wiki:WhichFiltertouse Which filter should one use?]
   * [wiki:EnsembleGeneration Generating ensembles]
 * Full Documentation
   * [wiki:CompilingPdaf Compiling the PDAF library]
   * [wiki:OnlineImplementationGuide_PDAF3 Implementation Guide for the online assimilation system]
   * [wiki:OfflineImplementationGuide_PDAF3 Implementation Guide for the offline assimilation program]
   * [wiki:DataAssimilationDiagnostics Diagnostic functions]
   * [wiki:PDAF_OMI_Overview_PDAF3 Observation handling and diagnostics using PDAF-OMI]
   * [wiki:ImplementGenerateObs_PDAF3 Generate synthetic observations for twin experiments]
   * [wiki:AvailableOptionsforInitPDAF Overview of filter-specific options]
   * [wiki:AddFilterAlgorithm_PDAF3 Adding a data assimilation method to PDAF]
   * [wiki:AuxiliaryRoutines Auxiliary and advanced functionality]
   * [wiki:RoutineOverviews Lists for direct access to documentation on particular routines]
   * [wiki:FAQ_PDAF FAQ]

 * [PublicationsandPresentations Publications and Presentations involving and about PDAF]
 * [wiki:CitingPDAF Citing PDAF]