 |
| Titel |
Emergent constraints on regional climate change projections |
| VerfasserIn |
Jonas Bhend, Penny Whetton |
| Konferenz |
EGU General Assembly 2014
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 16 (2014) |
| Datensatznummer |
250089524
|
| Publikation (Nr.) |
 EGU/EGU2014-3729.pdf |
|
|
|
|
|
| Zusammenfassung |
| While relying on output from multiple models for climate projections is standard practice,
there is no consensus on how best to combine simulations from a multimodel ensemble as of
yet. Here we apply the recently proposed "Second-order Exchangeability Analysis for
Multimodel Ensembles" by Rougier et al.to study the effect of recent observed regional
change as a constraint on future expected change. In this approach to combine multimodel
ensembles, the climate model outputs in the ensemble are assumed to be exchangeable
whereas the ensemble mean differs from the true climate system. Observations of the
climate system are then used to estimate this "ensemble discrepancy" and the updated
projections.
We analyse the relationship between recent trends and future projections in near-surface
temperature, rainfall and sea level pressure using simulations of the CMIP5 ensemble.
Preliminary results indicate that up to 40% of the variability in projected changes is related to
recent trends. At present, observed regional climate change therefore provides only weak
constraints for future projections. The greenhouse gas (GHG) component of recent change,
however, provides much stronger constraints. This highlights the differences in the
regional response to non-GHG forcing across different models and the need to
better understand the causes of observed regional change to find more relevant
constraints for regional projections. We validate the approach in a perfect model
framework using alternative models from the CMIP5 archive as pseudo-observations.
Finally, we compare projected regional changes with and without observational
constraints. In most regions we find little difference between the constrained and
unconstrained projections due to the weak relationship between recent and future
change and/or the consistency of observed recent change with simulated recent
changes. |
|
|
|
|
|
|