AMO-like Interdecadal Variability in the CMIP5 - Are Models Oversensitive to Prescribed Forcing?


Huang-Hsiung Hsu

15:55:00 - 16:45:00

101 , Mathematics Research Center Building (ori. New Math. Bldg.)

The Atlantic Multidecadal Oscillation (AMO), which fluctuates in a period of 60-70 years and has been found connected with many observed climate fluctuations, is one of the most explored interdecadal phenomena in recent years (e.g., Kushnir 1994, Schlesinger and Ramankutty 1994, Kerr 2000). Considering the significant impacts of the AMO on regional climate, which will be briefly reviewed in this presentation, whether the phenomenon is reasonably simulated by the current state-of-art climate model is an important issue for climate change research. Studies based on the third phase of Climate Model Intercomparison Project (CMIP3, Meehl et al. 2007) simulations found limited model capability (e.g., Kravtsov and Spannagle 2008, Knight 2009) in simulating AMO. This study revisited this issue by analyzing the CMIP5 data to firstly evaluate the model capability in the pre-industrial simulations. A reasonable model should simulate the AMO-like variability to a certain degree, if internal variability is the major reason for the existence of the AMO. Whether the AMO-like variability is enhanced or suppressed in the historical simulations were also examined through a comparison with the pre-industrial simulation. Both pre-industrial and historical simulations were started with the same initial condition. The former was simulated with constant CO2 concentration in 1850 and no any other observed forcing, while the latter was forced by prescribing "observed" (both natural and anthropogenic) forcings (Taylor et al. 2012 for details). The systematic differences between two sets of experiments were most likely due to the prescribed forcings in the historical simulations. In order to explore the model sensitivity to the prescribed forcing, interdecadal variability in the future projection is also briefly touched upon. Present study reveals the current status of our capability in simulating AMO-like interdecadal variability and provides further information for the interpretation of observed AMO variation in recent few decades.

For material related to this talk, click here.