 |
| Titel |
Perspectives of data assimilation for the climate of the Last Glacial Maximum |
| VerfasserIn |
Andre Paul, Martin Losch |
| Konferenz |
EGU General Assembly 2010
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 12 (2010) |
| Datensatznummer |
250044837
|
|
|
|
|
|
| Zusammenfassung |
| We are convinced that to further advance our understanding of the climate of the Last Glacial
Maximum (LGM, Â 19,000- 23,000 years before present) it is necessary to apply data
assimilation techniques to paleo-data and rigorously constrain the uncertainties of models as
well as data.
In the traditional (“forward”) method, models are tuned to reproduce observations by
adjusting a small number of parameters and repeating simulations in an ad-hoc
fashion. Although it would be possible to compute an explicit objective function,
this is rarely done. There are a number of “inverse” methods that use variational,
filtering and statistical techniques to overcome this crude tuning procedure and
combine paleo-data with a numerical model in a much more systematic way. In these
methods, the model (or even an ensemble of models) is integrated repeatedly and
variables that control the solution are adusted to minimize the departure from the
data.
For illustration, we used a classical energy balance climate model and applied the
so-called “adjoint method” to minimize the misfit between our model and sea-surface
temperature data for the Last Glacial Maximum (LGM, Â 19,000- 23,000 years before
present), taken from the Multiproxy Approach for the Reconstruction of the Glacial Ocean
surface (MARGO 2009). The “adjoint model” (derivative code) was generated by an “adjoint
compiler”. We optimized three parameters: the thermal diffusion coefficient, and two
constants in the parameterization of the outgoing longwave radiation that are related to the
sensitivity to changes in the zonal-mean surface temperature and the atmospheric CO2
concentration.
Using our simple example, we discuss the benefits as well as the issues of the adjoint
method. For example, by computing model sensitivities it may guide observational
efforts. However, it is still questionable whether the paleo-data available for the
LGM is accurate enough and its coverage sufficient to be useful for our purpose. |
|
|
|
|
|
|