Ensemble Data Assimilation: Algorithms and Software
Ensemble data assimilation is nowadays applied to various problems to estimate a model state and model parameters by combining the model predictions with observational data. At the Alfred Wegener Institute, the assimilation focuses on ocean-sea ice models and coupled ocean-biogeochemical models. The...
| Main Author: | |
|---|---|
| Format: | Conference Object |
| Language: | unknown |
| Published: |
2014
|
| Subjects: | |
| Online Access: | https://epic.awi.de/id/eprint/36550/ https://epic.awi.de/id/eprint/36550/1/Nerger_EnsDA_NMEFC_2014-10-10.pdf https://hdl.handle.net/10013/epic.44357 https://hdl.handle.net/10013/epic.44357.d001 |
| Summary: | Ensemble data assimilation is nowadays applied to various problems to estimate a model state and model parameters by combining the model predictions with observational data. At the Alfred Wegener Institute, the assimilation focuses on ocean-sea ice models and coupled ocean-biogeochemical models. The high dimension of realistic models requires particularly efficient algorithms that are also usable on supercomputers. For the application of such filters, the Parallel Data Assimilation Framework PDAF (http://pdaf.awi.de) was developed. PDAF provides an environment for ensemble integration as well as ensemble filters to assimilate observations. PDAF also provides synergies by allowing us to use the same data assimilation software for different models. The talk will provide an overview of developments in ensemble Kalman filters for high-dimensional models as well as on implementing ensemble data assimilation systems. |
|---|