 |
| Titel |
Parameter Identification in Ocean Models via Oneshot Optimization |
| VerfasserIn |
C. Kratzenstein, T. Slawig, A. Griewank |
| Konferenz |
EGU General Assembly 2012
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 14 (2012) |
| Datensatznummer |
250066626
|
|
|
|
|
|
| Zusammenfassung |
| We present an iterative method to solve data assimilation problems to identify climate model parameters, here applied to a 3-dimensional marine ecosystem model. The goal is to correct model parameters already during the simulation process in a piggyback manner. We augment the simulation tool with adjoint solvers to obtain derivatives and immediately use this sensitivity information to determine paramter corrections. We verify that the application of the Oneshot optimization strategy requires only a small multiple of the computational time and costs which are needed for one model spinup with fixed parameters. Numerical results are shown on the basis of a least-squares fit of model output of a nutrient-based 3D marine ecosystem model to given data. |
|
|
|
|
|