Read here. It is common knowledge that weather computer models are unreliable for accurate predictions that extend past a three day future. Climate models, which are closely related to weather models, suffer from the same non-reliable prediction issues as the longer than three day weather simulations.
In a state-of-the-art analysis, peer reviewed research by an IPCC climate expert has now confirmed that climate models are essentially worthless for predictions that stretch out a decade, which makes the models even worse than worthless for century-scale predictions.
Climate experts Latif and Keenlyside established the following:
1. "...examples of these internal [climate] variations are "the North Atlantic Oscillation (NAO), the El Niño/Southern Oscillation (ENSO), the Pacific Decadal Variability (PDV), and the Atlantic Multidecadal Variability (AMV)," all of which "project on global or hemispheric surface air temperature (SAT)...special attention to the variability of the Atlantic Meridional Overturning Circulation (AMOC), which they suggest is likely the origin of a considerable part of the decadal variability within the Atlantic Sector"
2. "...list numerous problems that hamper decadal climate predictability, among which is the fact that "the models suffer from large biases." In the cases of annual mean sea surface temperature (SST) and SAT over land, for example, they state that "typical errors can amount up to 10°C in certain regions,"..."
3. "...they add that several models also "fail to simulate a realistic El Niño/Southern Oscillation."..."
4. "...they indicate that "several assumptions have generally to be made about the process under consideration that cannot be rigorously justified, and this is a major source of uncertainty.""
5. "Another problem they discuss is the fact that "some components of the climate system are not well represented or not at all part of standard climate models," one example being the models' neglect of the stratosphere. This omission is quite serious, since Latif and Keenlyside say that "recent studies indicate that the mid-latitudinal response to both tropical and extra-tropical SST anomalies over the North Atlantic Sector may critically depend on stratospheric feedbacks,"..."
6. "An additional common model shortcoming, even in stand-alone integrations with models forced by observed SSTs, is that model simulations of rainfall in the Sahel "fail to reproduce the correct magnitude of the decadal precipitation anomalies.""
7. "Still another failure is the fact, as shown by Stroeve et al. (2007), that "virtually all climate models considerably underestimate the observed Arctic sea ice decline during the recent decades in the so-called 20th century integrations with prescribed (known natural and anthropogenic) observed forcing.""
8. "And added to these problems is the fact that "atmospheric chemistry and aerosol processes are still not well incorporated into current climate models.""
9. "...that "they [models] are incomplete and do not incorporate potentially important physics,""
10. "...that "the poor observational database does not allow a distinction between 'realistic' and 'unrealistic' simulations,"..."
11. "...state that "a sufficient understanding of the mechanisms of decadal-to-multidecadal variability is lacking," that "state-of-the-art climate models suffer from large biases,"..."
12. "...Therefore, they conclude that "it cannot be assumed that current climate models are well suited to realize the full decadal predictability potential,"..."
[Mojib Latif, Noel S. Keenlyside 2011: Deep Sea Research Part II: Topical Studies in Oceanography]
So, in summary, climate experts from the IPCC are conclusively stating that the IPCC and its climate experts have been wrong with their future climate predictions (scenarios) because the climate models are not "well suited" for doing so. This abysmal, comprehensive climate model failure thus applies to all past IPCC report predictions and will still be the case for future IPCC reports (AR5 and others).
Did we mention climate model predictions are worthless yet?
Previous climate-model and peer-reviewed postings.