 |
| Titel |
Robust Validation of ENSO in IPCC-Class Coupled Models |
| VerfasserIn |
Samantha Stevenson, Baylor Fox-Kemper, Markus Jochum |
| Konferenz |
EGU General Assembly 2010
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 12 (2010) |
| Datensatznummer |
250036397
|
|
|
|
|
|
| Zusammenfassung |
| Wavelet probability analysis, a new method of model validation, is used to assess the
performance of ENSO in a variety of coupled climate models. Wavelet probability analysis
relies on wavelet spectra for a given time series, for which the amount of spectral overlap
between subsets is measured using a quantity known as the wavelet probability index (WPI).
This approach provides quantitative estimates of model agreement relative to either
observations or other models, accompanied by well-defined confidence levels. ENSO, as
represented by the NINO3.4 index, has been examined in 2,000 year long coupled
integrations of both the new NCAR CCSM3.5 and GFDL’s CM2.1; interestingly, it is not
possible to distinguish either model from observations of NINO3.4 during 1949-2003, for
runs shorter than  200 years. At longer model run lengths, some inaccuracies are seen
in both CCSM3.5 and CM2.1 relative to observations. CCSM3.5 and CM2.1 are
compared to one another using hypothesis testing procedures, and changes in model
physics discussed in terms of their impact on ENSO. Finally, the method is applied to
non-equilibrium simulations, using both high-CO2 ‘ramp-up’ runs and selected IPCC
AR4 integrations. This allows the effect of changing CO2 levels on ENSO activity
to be examined, and the statistical significance of such effects to be determined. |
|
|
|
|
|
|